SpyBara
Go Premium

Reference 2026-09-28 22:57 UTC to 2026-09-29 22:57 UTC

33 files changed +1,086 −78. View all changes and history on the product overview
2026
Wed 30 22:59 Tue 29 22:57 Mon 28 22:57 Sun 27 23:58 Sat 26 23:59 Fri 25 23:58 Thu 24 23:58 Wed 23 23:58 Tue 22 23:57 Mon 21 23:00 Sat 19 23:00 Fri 18 22:59 Wed 16 22:58 Fri 11 20:00 Thu 10 18:01 Wed 9 16:02 Fri 4 23:59 Thu 3 23:00 Wed 2 01:58 Tue 1 01:57
Details

536 536 

537 - `"web_search"`537 - `"web_search"`

538 538 

539 - `ComputerUse object { include_screenshots, type }`

540 

541 Browser use in an OpenAI-hosted session.

542 

543 - `include_screenshots: boolean`

544 

545 Whether computer tool outputs include screenshots.

546 

547 - `type: "computer_use"`

548 

549 The type of the object. Always `computer_use`.

550 

551 - `"computer_use"`

552 

539 - `created_at: number`553 - `created_at: number`

540 554 

541 The Unix timestamp, in seconds, when the session was created.555 The Unix timestamp, in seconds, when the session was created.


554 568 

555 - `"none"`569 - `"none"`

556 570 

557 - `OpenAIHosted object { id, capability_directories, files, 5 more }`571 - `OpenAIHosted object { id, capability_directories, desktop, 7 more }`

558 572 

559 An environment hosted by OpenAI.573 An environment hosted by OpenAI.

560 574 


566 580 

567 Directories that contain capabilities exposed to the agent.581 Directories that contain capabilities exposed to the agent.

568 582 

583 - `desktop: object { enabled }`

584 

585 The effective desktop configuration.

586 

587 - `enabled: boolean`

588 

589 Whether the environment provisions a desktop and browser proxy.

590 

569 - `files: array of HostedEnvironmentFile`591 - `files: array of HostedEnvironmentFile`

570 592 

571 Files available in the environment, excluding their contents.593 Files available in the environment, excluding their contents.


636 658 

637 - `"restricted"`659 - `"restricted"`

638 660 

639 Allows access only to configured domains.661 Applies the configured domain restrictions.

640 662 

641 - `allowed_domains: array of string`663 - `allowed_domains: array of string`

642 664 


730 752 

731 - `"openai_hosted"`753 - `"openai_hosted"`

732 754 

755 - `container_size: optional "small" or "medium" or "large" or null`

756 

757 The effective CPU and memory tier, or null when unknown or outside the public tiers.

758 

759 - `"small"`

760 

761 - `"medium"`

762 

763 - `"large"`

764 

733 - `SelfHosted object { id, capability_directories, remote_url, 2 more }`765 - `SelfHosted object { id, capability_directories, remote_url, 2 more }`

734 766 

735 An environment hosted by the application.767 An environment hosted by the application.


774 806 

775 - `"agent.session"`807 - `"agent.session"`

776 808 

777 - `required_actions: array of object { arguments, call_id, name, 2 more } or object { environment_id, type }`809 - `required_actions: array of object { request, request_id, turn_id, type } or object { arguments, call_id, name, 2 more } or object { environment_id, type }`

778 810 

779 Actions that must be completed before the session can continue.811 Actions that must be completed before the session can continue.

780 812 

813 - `ComputerUseApprovalRequest object { request, request_id, turn_id, type }`

814 

815 Respond to a computer-use request.

816 

817 - `request: object { credential_origin, fields, options, 2 more } or object { origin, reason, type }`

818 

819 The information needed to render the request.

820 

821 - `BrowserAuthentication object { credential_origin, fields, options, 2 more }`

822 

823 A registered form awaiting the application's response.

824 

825 - `credential_origin: string or null`

826 

827 The registered form or frame origin where values will be entered.

828 

829 - `fields: array of object { id, label, required, type }`

830 

831 Controls to render. All submitted values are sensitive.

832 

833 - `id: string`

834 

835 The field ID to submit as field_id in a fields entry.

836 

837 - `label: string`

838 

839 The label to display beside the control.

840 

841 - `required: boolean`

842 

843 Whether this control requires a nonempty value.

844 

845 - `type: string`

846 

847 The rendering type, such as email, password, or text.

848 

849 - `options: array of object { id, field_ids, label }`

850 

851 Sign-in methods. Empty for a plain form.

852 

853 - `id: string`

854 

855 The option ID to submit as selected_option.

856 

857 - `field_ids: array of string`

858 

859 IDs from the registered fields that this method accepts.

860 

861 - `label: string`

862 

863 The method label to display.

864 

865 - `reason: string or null`

866 

867 Why the agent needs the user to sign in.

868 

869 - `type: "browser_authentication"`

870 

871 The type of the object. Always `browser_authentication`.

872 

873 - `"browser_authentication"`

874 

875 - `BrowserOriginAccess object { origin, reason, type }`

876 

877 A browser origin awaiting the application's approval decision.

878 

879 - `origin: string`

880 

881 The origin the browser needs permission to access.

882 

883 - `reason: string or null`

884 

885 The browser's explanation for this request, or null when unavailable.

886 

887 - `type: "browser_origin_access"`

888 

889 The type of the object. Always `browser_origin_access`.

890 

891 - `"browser_origin_access"`

892 

893 - `request_id: string`

894 

895 The registered request ID to echo when responding.

896 

897 - `turn_id: string`

898 

899 The turn that requested approval.

900 

901 - `type: "computer_use_approval_request"`

902 

903 The type of the object. Always `computer_use_approval_request`.

904 

905 - `"computer_use_approval_request"`

906 

781 - `FunctionCall object { arguments, call_id, name, 2 more }`907 - `FunctionCall object { arguments, call_id, name, 2 more }`

782 908 

783 Run a function tool and submit its result.909 Run a function tool and submit its result.


897 "status": "idle",1023 "status": "idle",

898 "required_actions": [1024 "required_actions": [

899 {1025 {

900 "type": "function_call",1026 "type": "computer_use_approval_request",

901 "turn_id": "turn_id",1027 "turn_id": "turn_id",

902 "call_id": "call_id",1028 "request_id": "request_id",

903 "name": "name",1029 "request": {

904 "arguments": {}1030 "type": "browser_authentication",

1031 "reason": "reason",

1032 "credential_origin": "credential_origin",

1033 "fields": [

1034 {

1035 "id": "id",

1036 "label": "label",

1037 "type": "type",

1038 "required": true

1039 }

1040 ],

1041 "options": [

1042 {

1043 "id": "id",

1044 "label": "label",

1045 "field_ids": [

1046 "string"

1047 ]

1048 }

1049 ]

1050 }

905 }1051 }

906 ],1052 ],

907 "error": "error",1053 "error": "error",


2628 2774 

2629 - `"mcp_call"`2775 - `"mcp_call"`

2630 2776 

2777 - `ComputerUseCall object { id, output, status, 3 more }`

2778 

2779 One execution of the platform-provided computer-use capability.

2780 

2781 - `id: string`

2782 

2783 The ID of the activity item.

2784 

2785 - `output: object { image_url, type } or null`

2786 

2787 The last screenshot emitted by the model. Null when screenshot inclusion is disabled or the call emitted no screenshot.

2788 

2789 - `image_url: string`

2790 

2791 The complete JPEG image as a base64 data URL.

2792 

2793 - `type: "computer_screenshot"`

2794 

2795 The content type. Always `computer_screenshot`.

2796 

2797 - `"computer_screenshot"`

2798 

2799 - `status: AgentFunctionCallStatus`

2800 

2801 The execution status of the activity.

2802 

2803 - `title: string or null`

2804 

2805 A model-generated description of the activity, when available.

2806 

2807 - `turn_id: string`

2808 

2809 The ID of the turn that contains this item.

2810 

2811 - `type: "computer_use_call"`

2812 

2813 The item type. Always `computer_use_call`.

2814 

2815 - `"computer_use_call"`

2816 

2817 - `ComputerUseApprovalRequest object { id, request, request_id, 2 more }`

2818 

2819 A credential-free history record of the emitted login request.

2820 

2821 - `id: string`

2822 

2823 The stable history item ID.

2824 

2825 - `request: object { credential_origin, fields, options, 2 more }`

2826 

2827 A registered form awaiting the application's response.

2828 

2829 - `credential_origin: string or null`

2830 

2831 The registered form or frame origin where values will be entered.

2832 

2833 - `fields: array of object { id, label, required, type }`

2834 

2835 Controls to render. All submitted values are sensitive.

2836 

2837 - `id: string`

2838 

2839 The field ID to submit as field_id in a fields entry.

2840 

2841 - `label: string`

2842 

2843 The label to display beside the control.

2844 

2845 - `required: boolean`

2846 

2847 Whether this control requires a nonempty value.

2848 

2849 - `type: string`

2850 

2851 The rendering type, such as email, password, or text.

2852 

2853 - `options: array of object { id, field_ids, label }`

2854 

2855 Sign-in methods. Empty for a plain form.

2856 

2857 - `id: string`

2858 

2859 The option ID to submit as selected_option.

2860 

2861 - `field_ids: array of string`

2862 

2863 IDs from the registered fields that this method accepts.

2864 

2865 - `label: string`

2866 

2867 The method label to display.

2868 

2869 - `reason: string or null`

2870 

2871 Why the agent needs the user to sign in.

2872 

2873 - `type: "browser_authentication"`

2874 

2875 The type of the object. Always `browser_authentication`.

2876 

2877 - `"browser_authentication"`

2878 

2879 - `request_id: string`

2880 

2881 - `turn_id: string`

2882 

2883 - `type: "computer_use_approval_request"`

2884 

2885 The item type. Always computer_use_approval_request.

2886 

2887 - `"computer_use_approval_request"`

2888 

2889 - `ComputerUseApprovalRequestResult object { id, request_id, response, 2 more }`

2890 

2891 A credential-free record of an admitted response, not proof of completion.

2892 

2893 - `id: string`

2894 

2895 The stable history item ID.

2896 

2897 - `request_id: string`

2898 

2899 The registered request answered by this item.

2900 

2901 - `response: object { action, selected_option, type } or object { action, type }`

2902 

2903 The admitted response, without submitted credential values.

2904 

2905 - `Submit object { action, selected_option, type }`

2906 

2907 - `action: "submit"`

2908 

2909 - `"submit"`

2910 

2911 - `selected_option: string or null`

2912 

2913 The chosen sign-in method, or null when no options were offered.

2914 

2915 - `type: "browser_authentication"`

2916 

2917 - `"browser_authentication"`

2918 

2919 - `Cancel object { action, type }`

2920 

2921 - `action: "cancel"`

2922 

2923 - `"cancel"`

2924 

2925 - `type: "browser_authentication"`

2926 

2927 - `"browser_authentication"`

2928 

2929 - `turn_id: string`

2930 

2931 The ID of the turn that contains this item.

2932 

2933 - `type: "computer_use_approval_request_result"`

2934 

2935 - `"computer_use_approval_request_result"`

2936 

2631 - `AgentWebSearchCallItem object { id, action, status, 2 more }`2937 - `AgentWebSearchCallItem object { id, action, status, 2 more }`

2632 2938 

2633 A web search call produced by the agent.2939 A web search call produced by the agent.


3350 3656 

3351 - `"web_search"`3657 - `"web_search"`

3352 3658 

3659 - `ComputerUse object { include_screenshots, type }`

3660 

3661 Browser use in an OpenAI-hosted session.

3662 

3663 - `include_screenshots: boolean`

3664 

3665 Whether computer tool outputs include screenshots.

3666 

3667 - `type: "computer_use"`

3668 

3669 The type of the object. Always `computer_use`.

3670 

3671 - `"computer_use"`

3672 

3353 - `created_at: number`3673 - `created_at: number`

3354 3674 

3355 The Unix timestamp, in seconds, when the session was created.3675 The Unix timestamp, in seconds, when the session was created.


3368 3688 

3369 - `"none"`3689 - `"none"`

3370 3690 

3371 - `OpenAIHosted object { id, capability_directories, files, 5 more }`3691 - `OpenAIHosted object { id, capability_directories, desktop, 7 more }`

3372 3692 

3373 An environment hosted by OpenAI.3693 An environment hosted by OpenAI.

3374 3694 


3380 3700 

3381 Directories that contain capabilities exposed to the agent.3701 Directories that contain capabilities exposed to the agent.

3382 3702 

3703 - `desktop: object { enabled }`

3704 

3705 The effective desktop configuration.

3706 

3707 - `enabled: boolean`

3708 

3709 Whether the environment provisions a desktop and browser proxy.

3710 

3383 - `files: array of HostedEnvironmentFile`3711 - `files: array of HostedEnvironmentFile`

3384 3712 

3385 Files available in the environment, excluding their contents.3713 Files available in the environment, excluding their contents.


3450 3778 

3451 - `"restricted"`3779 - `"restricted"`

3452 3780 

3453 Allows access only to configured domains.3781 Applies the configured domain restrictions.

3454 3782 

3455 - `allowed_domains: array of string`3783 - `allowed_domains: array of string`

3456 3784 


3544 3872 

3545 - `"openai_hosted"`3873 - `"openai_hosted"`

3546 3874 

3875 - `container_size: optional "small" or "medium" or "large" or null`

3876 

3877 The effective CPU and memory tier, or null when unknown or outside the public tiers.

3878 

3879 - `"small"`

3880 

3881 - `"medium"`

3882 

3883 - `"large"`

3884 

3547 - `SelfHosted object { id, capability_directories, remote_url, 2 more }`3885 - `SelfHosted object { id, capability_directories, remote_url, 2 more }`

3548 3886 

3549 An environment hosted by the application.3887 An environment hosted by the application.


3588 3926 

3589 - `"agent.session"`3927 - `"agent.session"`

3590 3928 

3591 - `required_actions: array of object { arguments, call_id, name, 2 more } or object { environment_id, type }`3929 - `required_actions: array of object { request, request_id, turn_id, type } or object { arguments, call_id, name, 2 more } or object { environment_id, type }`

3592 3930 

3593 Actions that must be completed before the session can continue.3931 Actions that must be completed before the session can continue.

3594 3932 

3933 - `ComputerUseApprovalRequest object { request, request_id, turn_id, type }`

3934 

3935 Respond to a computer-use request.

3936 

3937 - `request: object { credential_origin, fields, options, 2 more } or object { origin, reason, type }`

3938 

3939 The information needed to render the request.

3940 

3941 - `BrowserAuthentication object { credential_origin, fields, options, 2 more }`

3942 

3943 A registered form awaiting the application's response.

3944 

3945 - `credential_origin: string or null`

3946 

3947 The registered form or frame origin where values will be entered.

3948 

3949 - `fields: array of object { id, label, required, type }`

3950 

3951 Controls to render. All submitted values are sensitive.

3952 

3953 - `id: string`

3954 

3955 The field ID to submit as field_id in a fields entry.

3956 

3957 - `label: string`

3958 

3959 The label to display beside the control.

3960 

3961 - `required: boolean`

3962 

3963 Whether this control requires a nonempty value.

3964 

3965 - `type: string`

3966 

3967 The rendering type, such as email, password, or text.

3968 

3969 - `options: array of object { id, field_ids, label }`

3970 

3971 Sign-in methods. Empty for a plain form.

3972 

3973 - `id: string`

3974 

3975 The option ID to submit as selected_option.

3976 

3977 - `field_ids: array of string`

3978 

3979 IDs from the registered fields that this method accepts.

3980 

3981 - `label: string`

3982 

3983 The method label to display.

3984 

3985 - `reason: string or null`

3986 

3987 Why the agent needs the user to sign in.

3988 

3989 - `type: "browser_authentication"`

3990 

3991 The type of the object. Always `browser_authentication`.

3992 

3993 - `"browser_authentication"`

3994 

3995 - `BrowserOriginAccess object { origin, reason, type }`

3996 

3997 A browser origin awaiting the application's approval decision.

3998 

3999 - `origin: string`

4000 

4001 The origin the browser needs permission to access.

4002 

4003 - `reason: string or null`

4004 

4005 The browser's explanation for this request, or null when unavailable.

4006 

4007 - `type: "browser_origin_access"`

4008 

4009 The type of the object. Always `browser_origin_access`.

4010 

4011 - `"browser_origin_access"`

4012 

4013 - `request_id: string`

4014 

4015 The registered request ID to echo when responding.

4016 

4017 - `turn_id: string`

4018 

4019 The turn that requested approval.

4020 

4021 - `type: "computer_use_approval_request"`

4022 

4023 The type of the object. Always `computer_use_approval_request`.

4024 

4025 - `"computer_use_approval_request"`

4026 

3595 - `FunctionCall object { arguments, call_id, name, 2 more }`4027 - `FunctionCall object { arguments, call_id, name, 2 more }`

3596 4028 

3597 Run a function tool and submit its result.4029 Run a function tool and submit its result.


3711 "status": "idle",4143 "status": "idle",

3712 "required_actions": [4144 "required_actions": [

3713 {4145 {

3714 "type": "function_call",4146 "type": "computer_use_approval_request",

3715 "turn_id": "turn_id",4147 "turn_id": "turn_id",

3716 "call_id": "call_id",4148 "request_id": "request_id",

3717 "name": "name",4149 "request": {

3718 "arguments": {}4150 "type": "browser_authentication",

4151 "reason": "reason",

4152 "credential_origin": "credential_origin",

4153 "fields": [

4154 {

4155 "id": "id",

4156 "label": "label",

4157 "type": "type",

4158 "required": true

4159 }

4160 ],

4161 "options": [

4162 {

4163 "id": "id",

4164 "label": "label",

4165 "field_ids": [

4166 "string"

4167 ]

4168 }

4169 ]

4170 }

3719 }4171 }

3720 ],4172 ],

3721 "error": "error",4173 "error": "error",


4102 4554 

4103 - `"web_search"`4555 - `"web_search"`

4104 4556 

4557 - `ComputerUse object { include_screenshots, type }`

4558 

4559 Browser use in an OpenAI-hosted session.

4560 

4561 - `include_screenshots: boolean`

4562 

4563 Whether computer tool outputs include screenshots.

4564 

4565 - `type: "computer_use"`

4566 

4567 The type of the object. Always `computer_use`.

4568 

4569 - `"computer_use"`

4570 

4105 - `created_at: number`4571 - `created_at: number`

4106 4572 

4107 The Unix timestamp, in seconds, when the session was created.4573 The Unix timestamp, in seconds, when the session was created.


4120 4586 

4121 - `"none"`4587 - `"none"`

4122 4588 

4123 - `OpenAIHosted object { id, capability_directories, files, 5 more }`4589 - `OpenAIHosted object { id, capability_directories, desktop, 7 more }`

4124 4590 

4125 An environment hosted by OpenAI.4591 An environment hosted by OpenAI.

4126 4592 


4132 4598 

4133 Directories that contain capabilities exposed to the agent.4599 Directories that contain capabilities exposed to the agent.

4134 4600 

4601 - `desktop: object { enabled }`

4602 

4603 The effective desktop configuration.

4604 

4605 - `enabled: boolean`

4606 

4607 Whether the environment provisions a desktop and browser proxy.

4608 

4135 - `files: array of HostedEnvironmentFile`4609 - `files: array of HostedEnvironmentFile`

4136 4610 

4137 Files available in the environment, excluding their contents.4611 Files available in the environment, excluding their contents.


4202 4676 

4203 - `"restricted"`4677 - `"restricted"`

4204 4678 

4205 Allows access only to configured domains.4679 Applies the configured domain restrictions.

4206 4680 

4207 - `allowed_domains: array of string`4681 - `allowed_domains: array of string`

4208 4682 


4296 4770 

4297 - `"openai_hosted"`4771 - `"openai_hosted"`

4298 4772 

4773 - `container_size: optional "small" or "medium" or "large" or null`

4774 

4775 The effective CPU and memory tier, or null when unknown or outside the public tiers.

4776 

4777 - `"small"`

4778 

4779 - `"medium"`

4780 

4781 - `"large"`

4782 

4299 - `SelfHosted object { id, capability_directories, remote_url, 2 more }`4783 - `SelfHosted object { id, capability_directories, remote_url, 2 more }`

4300 4784 

4301 An environment hosted by the application.4785 An environment hosted by the application.


4340 4824 

4341 - `"agent.session"`4825 - `"agent.session"`

4342 4826 

4343 - `required_actions: array of object { arguments, call_id, name, 2 more } or object { environment_id, type }`4827 - `required_actions: array of object { request, request_id, turn_id, type } or object { arguments, call_id, name, 2 more } or object { environment_id, type }`

4344 4828 

4345 Actions that must be completed before the session can continue.4829 Actions that must be completed before the session can continue.

4346 4830 

4831 - `ComputerUseApprovalRequest object { request, request_id, turn_id, type }`

4832 

4833 Respond to a computer-use request.

4834 

4835 - `request: object { credential_origin, fields, options, 2 more } or object { origin, reason, type }`

4836 

4837 The information needed to render the request.

4838 

4839 - `BrowserAuthentication object { credential_origin, fields, options, 2 more }`

4840 

4841 A registered form awaiting the application's response.

4842 

4843 - `credential_origin: string or null`

4844 

4845 The registered form or frame origin where values will be entered.

4846 

4847 - `fields: array of object { id, label, required, type }`

4848 

4849 Controls to render. All submitted values are sensitive.

4850 

4851 - `id: string`

4852 

4853 The field ID to submit as field_id in a fields entry.

4854 

4855 - `label: string`

4856 

4857 The label to display beside the control.

4858 

4859 - `required: boolean`

4860 

4861 Whether this control requires a nonempty value.

4862 

4863 - `type: string`

4864 

4865 The rendering type, such as email, password, or text.

4866 

4867 - `options: array of object { id, field_ids, label }`

4868 

4869 Sign-in methods. Empty for a plain form.

4870 

4871 - `id: string`

4872 

4873 The option ID to submit as selected_option.

4874 

4875 - `field_ids: array of string`

4876 

4877 IDs from the registered fields that this method accepts.

4878 

4879 - `label: string`

4880 

4881 The method label to display.

4882 

4883 - `reason: string or null`

4884 

4885 Why the agent needs the user to sign in.

4886 

4887 - `type: "browser_authentication"`

4888 

4889 The type of the object. Always `browser_authentication`.

4890 

4891 - `"browser_authentication"`

4892 

4893 - `BrowserOriginAccess object { origin, reason, type }`

4894 

4895 A browser origin awaiting the application's approval decision.

4896 

4897 - `origin: string`

4898 

4899 The origin the browser needs permission to access.

4900 

4901 - `reason: string or null`

4902 

4903 The browser's explanation for this request, or null when unavailable.

4904 

4905 - `type: "browser_origin_access"`

4906 

4907 The type of the object. Always `browser_origin_access`.

4908 

4909 - `"browser_origin_access"`

4910 

4911 - `request_id: string`

4912 

4913 The registered request ID to echo when responding.

4914 

4915 - `turn_id: string`

4916 

4917 The turn that requested approval.

4918 

4919 - `type: "computer_use_approval_request"`

4920 

4921 The type of the object. Always `computer_use_approval_request`.

4922 

4923 - `"computer_use_approval_request"`

4924 

4347 - `FunctionCall object { arguments, call_id, name, 2 more }`4925 - `FunctionCall object { arguments, call_id, name, 2 more }`

4348 4926 

4349 Run a function tool and submit its result.4927 Run a function tool and submit its result.


4463 "status": "idle",5041 "status": "idle",

4464 "required_actions": [5042 "required_actions": [

4465 {5043 {

4466 "type": "function_call",5044 "type": "computer_use_approval_request",

4467 "turn_id": "turn_id",5045 "turn_id": "turn_id",

4468 "call_id": "call_id",5046 "request_id": "request_id",

4469 "name": "name",5047 "request": {

4470 "arguments": {}5048 "type": "browser_authentication",

5049 "reason": "reason",

5050 "credential_origin": "credential_origin",

5051 "fields": [

5052 {

5053 "id": "id",

5054 "label": "label",

5055 "type": "type",

5056 "required": true

5057 }

5058 ],

5059 "options": [

5060 {

5061 "id": "id",

5062 "label": "label",

5063 "field_ids": [

5064 "string"

5065 ]

5066 }

5067 ]

5068 }

4471 }5069 }

4472 ],5070 ],

4473 "error": "error",5071 "error": "error",


4854 5452 

4855 - `"web_search"`5453 - `"web_search"`

4856 5454 

5455 - `ComputerUse object { include_screenshots, type }`

5456 

5457 Browser use in an OpenAI-hosted session.

5458 

5459 - `include_screenshots: boolean`

5460 

5461 Whether computer tool outputs include screenshots.

5462 

5463 - `type: "computer_use"`

5464 

5465 The type of the object. Always `computer_use`.

5466 

5467 - `"computer_use"`

5468 

4857 - `created_at: number`5469 - `created_at: number`

4858 5470 

4859 The Unix timestamp, in seconds, when the session was created.5471 The Unix timestamp, in seconds, when the session was created.


4872 5484 

4873 - `"none"`5485 - `"none"`

4874 5486 

4875 - `OpenAIHosted object { id, capability_directories, files, 5 more }`5487 - `OpenAIHosted object { id, capability_directories, desktop, 7 more }`

4876 5488 

4877 An environment hosted by OpenAI.5489 An environment hosted by OpenAI.

4878 5490 


4884 5496 

4885 Directories that contain capabilities exposed to the agent.5497 Directories that contain capabilities exposed to the agent.

4886 5498 

5499 - `desktop: object { enabled }`

5500 

5501 The effective desktop configuration.

5502 

5503 - `enabled: boolean`

5504 

5505 Whether the environment provisions a desktop and browser proxy.

5506 

4887 - `files: array of HostedEnvironmentFile`5507 - `files: array of HostedEnvironmentFile`

4888 5508 

4889 Files available in the environment, excluding their contents.5509 Files available in the environment, excluding their contents.


4954 5574 

4955 - `"restricted"`5575 - `"restricted"`

4956 5576 

4957 Allows access only to configured domains.5577 Applies the configured domain restrictions.

4958 5578 

4959 - `allowed_domains: array of string`5579 - `allowed_domains: array of string`

4960 5580 


5048 5668 

5049 - `"openai_hosted"`5669 - `"openai_hosted"`

5050 5670 

5671 - `container_size: optional "small" or "medium" or "large" or null`

5672 

5673 The effective CPU and memory tier, or null when unknown or outside the public tiers.

5674 

5675 - `"small"`

5676 

5677 - `"medium"`

5678 

5679 - `"large"`

5680 

5051 - `SelfHosted object { id, capability_directories, remote_url, 2 more }`5681 - `SelfHosted object { id, capability_directories, remote_url, 2 more }`

5052 5682 

5053 An environment hosted by the application.5683 An environment hosted by the application.


5092 5722 

5093 - `"agent.session"`5723 - `"agent.session"`

5094 5724 

5095 - `required_actions: array of object { arguments, call_id, name, 2 more } or object { environment_id, type }`5725 - `required_actions: array of object { request, request_id, turn_id, type } or object { arguments, call_id, name, 2 more } or object { environment_id, type }`

5096 5726 

5097 Actions that must be completed before the session can continue.5727 Actions that must be completed before the session can continue.

5098 5728 

5729 - `ComputerUseApprovalRequest object { request, request_id, turn_id, type }`

5730 

5731 Respond to a computer-use request.

5732 

5733 - `request: object { credential_origin, fields, options, 2 more } or object { origin, reason, type }`

5734 

5735 The information needed to render the request.

5736 

5737 - `BrowserAuthentication object { credential_origin, fields, options, 2 more }`

5738 

5739 A registered form awaiting the application's response.

5740 

5741 - `credential_origin: string or null`

5742 

5743 The registered form or frame origin where values will be entered.

5744 

5745 - `fields: array of object { id, label, required, type }`

5746 

5747 Controls to render. All submitted values are sensitive.

5748 

5749 - `id: string`

5750 

5751 The field ID to submit as field_id in a fields entry.

5752 

5753 - `label: string`

5754 

5755 The label to display beside the control.

5756 

5757 - `required: boolean`

5758 

5759 Whether this control requires a nonempty value.

5760 

5761 - `type: string`

5762 

5763 The rendering type, such as email, password, or text.

5764 

5765 - `options: array of object { id, field_ids, label }`

5766 

5767 Sign-in methods. Empty for a plain form.

5768 

5769 - `id: string`

5770 

5771 The option ID to submit as selected_option.

5772 

5773 - `field_ids: array of string`

5774 

5775 IDs from the registered fields that this method accepts.

5776 

5777 - `label: string`

5778 

5779 The method label to display.

5780 

5781 - `reason: string or null`

5782 

5783 Why the agent needs the user to sign in.

5784 

5785 - `type: "browser_authentication"`

5786 

5787 The type of the object. Always `browser_authentication`.

5788 

5789 - `"browser_authentication"`

5790 

5791 - `BrowserOriginAccess object { origin, reason, type }`

5792 

5793 A browser origin awaiting the application's approval decision.

5794 

5795 - `origin: string`

5796 

5797 The origin the browser needs permission to access.

5798 

5799 - `reason: string or null`

5800 

5801 The browser's explanation for this request, or null when unavailable.

5802 

5803 - `type: "browser_origin_access"`

5804 

5805 The type of the object. Always `browser_origin_access`.

5806 

5807 - `"browser_origin_access"`

5808 

5809 - `request_id: string`

5810 

5811 The registered request ID to echo when responding.

5812 

5813 - `turn_id: string`

5814 

5815 The turn that requested approval.

5816 

5817 - `type: "computer_use_approval_request"`

5818 

5819 The type of the object. Always `computer_use_approval_request`.

5820 

5821 - `"computer_use_approval_request"`

5822 

5099 - `FunctionCall object { arguments, call_id, name, 2 more }`5823 - `FunctionCall object { arguments, call_id, name, 2 more }`

5100 5824 

5101 Run a function tool and submit its result.5825 Run a function tool and submit its result.


5215 "status": "idle",5939 "status": "idle",

5216 "required_actions": [5940 "required_actions": [

5217 {5941 {

5218 "type": "function_call",5942 "type": "computer_use_approval_request",

5219 "turn_id": "turn_id",5943 "turn_id": "turn_id",

5220 "call_id": "call_id",5944 "request_id": "request_id",

5221 "name": "name",5945 "request": {

5222 "arguments": {}5946 "type": "browser_authentication",

5947 "reason": "reason",

5948 "credential_origin": "credential_origin",

5949 "fields": [

5950 {

5951 "id": "id",

5952 "label": "label",

5953 "type": "type",

5954 "required": true

5955 }

5956 ],

5957 "options": [

5958 {

5959 "id": "id",

5960 "label": "label",

5961 "field_ids": [

5962 "string"

5963 ]

5964 }

5965 ]

5966 }

5223 }5967 }

5224 ],5968 ],

5225 "error": "error",5969 "error": "error",


5606 6350 

5607 - `"web_search"`6351 - `"web_search"`

5608 6352 

6353 - `ComputerUse object { include_screenshots, type }`

6354 

6355 Browser use in an OpenAI-hosted session.

6356 

6357 - `include_screenshots: boolean`

6358 

6359 Whether computer tool outputs include screenshots.

6360 

6361 - `type: "computer_use"`

6362 

6363 The type of the object. Always `computer_use`.

6364 

6365 - `"computer_use"`

6366 

5609 - `created_at: number`6367 - `created_at: number`

5610 6368 

5611 The Unix timestamp, in seconds, when the session was created.6369 The Unix timestamp, in seconds, when the session was created.


5624 6382 

5625 - `"none"`6383 - `"none"`

5626 6384 

5627 - `OpenAIHosted object { id, capability_directories, files, 5 more }`6385 - `OpenAIHosted object { id, capability_directories, desktop, 7 more }`

5628 6386 

5629 An environment hosted by OpenAI.6387 An environment hosted by OpenAI.

5630 6388 


5636 6394 

5637 Directories that contain capabilities exposed to the agent.6395 Directories that contain capabilities exposed to the agent.

5638 6396 

6397 - `desktop: object { enabled }`

6398 

6399 The effective desktop configuration.

6400 

6401 - `enabled: boolean`

6402 

6403 Whether the environment provisions a desktop and browser proxy.

6404 

5639 - `files: array of HostedEnvironmentFile`6405 - `files: array of HostedEnvironmentFile`

5640 6406 

5641 Files available in the environment, excluding their contents.6407 Files available in the environment, excluding their contents.


5706 6472 

5707 - `"restricted"`6473 - `"restricted"`

5708 6474 

5709 Allows access only to configured domains.6475 Applies the configured domain restrictions.

5710 6476 

5711 - `allowed_domains: array of string`6477 - `allowed_domains: array of string`

5712 6478 


5800 6566 

5801 - `"openai_hosted"`6567 - `"openai_hosted"`

5802 6568 

6569 - `container_size: optional "small" or "medium" or "large" or null`

6570 

6571 The effective CPU and memory tier, or null when unknown or outside the public tiers.

6572 

6573 - `"small"`

6574 

6575 - `"medium"`

6576 

6577 - `"large"`

6578 

5803 - `SelfHosted object { id, capability_directories, remote_url, 2 more }`6579 - `SelfHosted object { id, capability_directories, remote_url, 2 more }`

5804 6580 

5805 An environment hosted by the application.6581 An environment hosted by the application.


5844 6620 

5845 - `"agent.session"`6621 - `"agent.session"`

5846 6622 

5847 - `required_actions: array of object { arguments, call_id, name, 2 more } or object { environment_id, type }`6623 - `required_actions: array of object { request, request_id, turn_id, type } or object { arguments, call_id, name, 2 more } or object { environment_id, type }`

5848 6624 

5849 Actions that must be completed before the session can continue.6625 Actions that must be completed before the session can continue.

5850 6626 

6627 - `ComputerUseApprovalRequest object { request, request_id, turn_id, type }`

6628 

6629 Respond to a computer-use request.

6630 

6631 - `request: object { credential_origin, fields, options, 2 more } or object { origin, reason, type }`

6632 

6633 The information needed to render the request.

6634 

6635 - `BrowserAuthentication object { credential_origin, fields, options, 2 more }`

6636 

6637 A registered form awaiting the application's response.

6638 

6639 - `credential_origin: string or null`

6640 

6641 The registered form or frame origin where values will be entered.

6642 

6643 - `fields: array of object { id, label, required, type }`

6644 

6645 Controls to render. All submitted values are sensitive.

6646 

6647 - `id: string`

6648 

6649 The field ID to submit as field_id in a fields entry.

6650 

6651 - `label: string`

6652 

6653 The label to display beside the control.

6654 

6655 - `required: boolean`

6656 

6657 Whether this control requires a nonempty value.

6658 

6659 - `type: string`

6660 

6661 The rendering type, such as email, password, or text.

6662 

6663 - `options: array of object { id, field_ids, label }`

6664 

6665 Sign-in methods. Empty for a plain form.

6666 

6667 - `id: string`

6668 

6669 The option ID to submit as selected_option.

6670 

6671 - `field_ids: array of string`

6672 

6673 IDs from the registered fields that this method accepts.

6674 

6675 - `label: string`

6676 

6677 The method label to display.

6678 

6679 - `reason: string or null`

6680 

6681 Why the agent needs the user to sign in.

6682 

6683 - `type: "browser_authentication"`

6684 

6685 The type of the object. Always `browser_authentication`.

6686 

6687 - `"browser_authentication"`

6688 

6689 - `BrowserOriginAccess object { origin, reason, type }`

6690 

6691 A browser origin awaiting the application's approval decision.

6692 

6693 - `origin: string`

6694 

6695 The origin the browser needs permission to access.

6696 

6697 - `reason: string or null`

6698 

6699 The browser's explanation for this request, or null when unavailable.

6700 

6701 - `type: "browser_origin_access"`

6702 

6703 The type of the object. Always `browser_origin_access`.

6704 

6705 - `"browser_origin_access"`

6706 

6707 - `request_id: string`

6708 

6709 The registered request ID to echo when responding.

6710 

6711 - `turn_id: string`

6712 

6713 The turn that requested approval.

6714 

6715 - `type: "computer_use_approval_request"`

6716 

6717 The type of the object. Always `computer_use_approval_request`.

6718 

6719 - `"computer_use_approval_request"`

6720 

5851 - `FunctionCall object { arguments, call_id, name, 2 more }`6721 - `FunctionCall object { arguments, call_id, name, 2 more }`

5852 6722 

5853 Run a function tool and submit its result.6723 Run a function tool and submit its result.


5967 "status": "idle",6837 "status": "idle",

5968 "required_actions": [6838 "required_actions": [

5969 {6839 {

5970 "type": "function_call",6840 "type": "computer_use_approval_request",

5971 "turn_id": "turn_id",6841 "turn_id": "turn_id",

5972 "call_id": "call_id",6842 "request_id": "request_id",

5973 "name": "name",6843 "request": {

5974 "arguments": {}6844 "type": "browser_authentication",

6845 "reason": "reason",

6846 "credential_origin": "credential_origin",

6847 "fields": [

6848 {

6849 "id": "id",

6850 "label": "label",

6851 "type": "type",

6852 "required": true

6853 }

6854 ],

6855 "options": [

6856 {

6857 "id": "id",

6858 "label": "label",

6859 "field_ids": [

6860 "string"

6861 ]

6862 }

6863 ]

6864 }

5975 }6865 }

5976 ],6866 ],

5977 "error": "error",6867 "error": "error",


7013 7903 

7014 - `"mcp_call"`7904 - `"mcp_call"`

7015 7905 

7906 - `ComputerUseCall object { id, output, status, 3 more }`

7907 

7908 One execution of the platform-provided computer-use capability.

7909 

7910 - `id: string`

7911 

7912 The ID of the activity item.

7913 

7914 - `output: object { image_url, type } or null`

7915 

7916 The last screenshot emitted by the model. Null when screenshot inclusion is disabled or the call emitted no screenshot.

7917 

7918 - `image_url: string`

7919 

7920 The complete JPEG image as a base64 data URL.

7921 

7922 - `type: "computer_screenshot"`

7923 

7924 The content type. Always `computer_screenshot`.

7925 

7926 - `"computer_screenshot"`

7927 

7928 - `status: AgentFunctionCallStatus`

7929 

7930 The execution status of the activity.

7931 

7932 - `title: string or null`

7933 

7934 A model-generated description of the activity, when available.

7935 

7936 - `turn_id: string`

7937 

7938 The ID of the turn that contains this item.

7939 

7940 - `type: "computer_use_call"`

7941 

7942 The item type. Always `computer_use_call`.

7943 

7944 - `"computer_use_call"`

7945 

7946 - `ComputerUseApprovalRequest object { id, request, request_id, 2 more }`

7947 

7948 A credential-free history record of the emitted login request.

7949 

7950 - `id: string`

7951 

7952 The stable history item ID.

7953 

7954 - `request: object { credential_origin, fields, options, 2 more }`

7955 

7956 A registered form awaiting the application's response.

7957 

7958 - `credential_origin: string or null`

7959 

7960 The registered form or frame origin where values will be entered.

7961 

7962 - `fields: array of object { id, label, required, type }`

7963 

7964 Controls to render. All submitted values are sensitive.

7965 

7966 - `id: string`

7967 

7968 The field ID to submit as field_id in a fields entry.

7969 

7970 - `label: string`

7971 

7972 The label to display beside the control.

7973 

7974 - `required: boolean`

7975 

7976 Whether this control requires a nonempty value.

7977 

7978 - `type: string`

7979 

7980 The rendering type, such as email, password, or text.

7981 

7982 - `options: array of object { id, field_ids, label }`

7983 

7984 Sign-in methods. Empty for a plain form.

7985 

7986 - `id: string`

7987 

7988 The option ID to submit as selected_option.

7989 

7990 - `field_ids: array of string`

7991 

7992 IDs from the registered fields that this method accepts.

7993 

7994 - `label: string`

7995 

7996 The method label to display.

7997 

7998 - `reason: string or null`

7999 

8000 Why the agent needs the user to sign in.

8001 

8002 - `type: "browser_authentication"`

8003 

8004 The type of the object. Always `browser_authentication`.

8005 

8006 - `"browser_authentication"`

8007 

8008 - `request_id: string`

8009 

8010 - `turn_id: string`

8011 

8012 - `type: "computer_use_approval_request"`

8013 

8014 The item type. Always computer_use_approval_request.

8015 

8016 - `"computer_use_approval_request"`

8017 

7016 - `AgentWebSearchCallItem object { id, action, status, 2 more }`8018 - `AgentWebSearchCallItem object { id, action, status, 2 more }`

7017 8019 

7018 A web search call produced by the agent.8020 A web search call produced by the agent.

Details

11 11 

12Use this guide to migrate your integration to the [Responses API](https://developers.openai.com/api/docs/guides/migrate-to-responses).12Use this guide to migrate your integration to the [Responses API](https://developers.openai.com/api/docs/guides/migrate-to-responses).

13 13 

14Responses are simpler—send input items and get output items back. With the Responses API, you also get better performance and new features like [deep research](https://developers.openai.com/api/docs/guides/deep-research), [MCP](https://developers.openai.com/api/docs/guides/tools-connectors-mcp), and [computer use](https://developers.openai.com/api/docs/guides/tools-computer-use). This change also lets you manage conversations instead of passing back `previous_response_id`.14Responses are simpler—send input items and get output items back. With the Responses API, you also get better performance and new features like [web search](https://developers.openai.com/api/docs/guides/tools-web-search), [MCP](https://developers.openai.com/api/docs/guides/tools-connectors-mcp), and [computer use](https://developers.openai.com/api/docs/guides/tools-computer-use). This change also lets you manage conversations instead of passing back `previous_response_id`.

15 15 

16### What's changed?16### What's changed?

17 17 

Details

11 11 

12Use this guide to migrate your integration to the [Responses API](https://developers.openai.com/api/docs/guides/migrate-to-responses).12Use this guide to migrate your integration to the [Responses API](https://developers.openai.com/api/docs/guides/migrate-to-responses).

13 13 

14Responses are simpler—send input items and get output items back. With the Responses API, you also get better performance and new features like [deep research](https://developers.openai.com/api/docs/guides/deep-research), [MCP](https://developers.openai.com/api/docs/guides/tools-connectors-mcp), and [computer use](https://developers.openai.com/api/docs/guides/tools-computer-use). This change also lets you manage conversations instead of passing back `previous_response_id`.14Responses are simpler—send input items and get output items back. With the Responses API, you also get better performance and new features like [web search](https://developers.openai.com/api/docs/guides/tools-web-search), [MCP](https://developers.openai.com/api/docs/guides/tools-connectors-mcp), and [computer use](https://developers.openai.com/api/docs/guides/tools-computer-use). This change also lets you manage conversations instead of passing back `previous_response_id`.

15 15 

16### What's changed?16### What's changed?

17 17 

Details

11 11 

12Use this guide to migrate your integration to the [Responses API](https://developers.openai.com/api/docs/guides/migrate-to-responses).12Use this guide to migrate your integration to the [Responses API](https://developers.openai.com/api/docs/guides/migrate-to-responses).

13 13 

14Responses are simpler—send input items and get output items back. With the Responses API, you also get better performance and new features like [deep research](https://developers.openai.com/api/docs/guides/deep-research), [MCP](https://developers.openai.com/api/docs/guides/tools-connectors-mcp), and [computer use](https://developers.openai.com/api/docs/guides/tools-computer-use). This change also lets you manage conversations instead of passing back `previous_response_id`.14Responses are simpler—send input items and get output items back. With the Responses API, you also get better performance and new features like [web search](https://developers.openai.com/api/docs/guides/tools-web-search), [MCP](https://developers.openai.com/api/docs/guides/tools-connectors-mcp), and [computer use](https://developers.openai.com/api/docs/guides/tools-computer-use). This change also lets you manage conversations instead of passing back `previous_response_id`.

15 15 

16### What's changed?16### What's changed?

17 17 

Details

11 11 

12Use this guide to migrate your integration to the [Responses API](https://developers.openai.com/api/docs/guides/migrate-to-responses).12Use this guide to migrate your integration to the [Responses API](https://developers.openai.com/api/docs/guides/migrate-to-responses).

13 13 

14Responses are simpler—send input items and get output items back. With the Responses API, you also get better performance and new features like [deep research](https://developers.openai.com/api/docs/guides/deep-research), [MCP](https://developers.openai.com/api/docs/guides/tools-connectors-mcp), and [computer use](https://developers.openai.com/api/docs/guides/tools-computer-use). This change also lets you manage conversations instead of passing back `previous_response_id`.14Responses are simpler—send input items and get output items back. With the Responses API, you also get better performance and new features like [web search](https://developers.openai.com/api/docs/guides/tools-web-search), [MCP](https://developers.openai.com/api/docs/guides/tools-connectors-mcp), and [computer use](https://developers.openai.com/api/docs/guides/tools-computer-use). This change also lets you manage conversations instead of passing back `previous_response_id`.

15 15 

16### What's changed?16### What's changed?

17 17 

Details

11 11 

12Use this guide to migrate your integration to the [Responses API](https://developers.openai.com/api/docs/guides/migrate-to-responses).12Use this guide to migrate your integration to the [Responses API](https://developers.openai.com/api/docs/guides/migrate-to-responses).

13 13 

14Responses are simpler—send input items and get output items back. With the Responses API, you also get better performance and new features like [deep research](https://developers.openai.com/api/docs/guides/deep-research), [MCP](https://developers.openai.com/api/docs/guides/tools-connectors-mcp), and [computer use](https://developers.openai.com/api/docs/guides/tools-computer-use). This change also lets you manage conversations instead of passing back `previous_response_id`.14Responses are simpler—send input items and get output items back. With the Responses API, you also get better performance and new features like [web search](https://developers.openai.com/api/docs/guides/tools-web-search), [MCP](https://developers.openai.com/api/docs/guides/tools-connectors-mcp), and [computer use](https://developers.openai.com/api/docs/guides/tools-computer-use). This change also lets you manage conversations instead of passing back `previous_response_id`.

15 15 

16### What's changed?16### What's changed?

17 17 

Details

11 11 

12Use this guide to migrate your integration to the [Responses API](https://developers.openai.com/api/docs/guides/migrate-to-responses).12Use this guide to migrate your integration to the [Responses API](https://developers.openai.com/api/docs/guides/migrate-to-responses).

13 13 

14Responses are simpler—send input items and get output items back. With the Responses API, you also get better performance and new features like [deep research](https://developers.openai.com/api/docs/guides/deep-research), [MCP](https://developers.openai.com/api/docs/guides/tools-connectors-mcp), and [computer use](https://developers.openai.com/api/docs/guides/tools-computer-use). This change also lets you manage conversations instead of passing back `previous_response_id`.14Responses are simpler—send input items and get output items back. With the Responses API, you also get better performance and new features like [web search](https://developers.openai.com/api/docs/guides/tools-web-search), [MCP](https://developers.openai.com/api/docs/guides/tools-connectors-mcp), and [computer use](https://developers.openai.com/api/docs/guides/tools-computer-use). This change also lets you manage conversations instead of passing back `previous_response_id`.

15 15 

16### What's changed?16### What's changed?

17 17 

Details

11 11 

12Use this guide to migrate your integration to the [Responses API](https://developers.openai.com/api/docs/guides/migrate-to-responses).12Use this guide to migrate your integration to the [Responses API](https://developers.openai.com/api/docs/guides/migrate-to-responses).

13 13 

14Responses are simpler—send input items and get output items back. With the Responses API, you also get better performance and new features like [deep research](https://developers.openai.com/api/docs/guides/deep-research), [MCP](https://developers.openai.com/api/docs/guides/tools-connectors-mcp), and [computer use](https://developers.openai.com/api/docs/guides/tools-computer-use). This change also lets you manage conversations instead of passing back `previous_response_id`.14Responses are simpler—send input items and get output items back. With the Responses API, you also get better performance and new features like [web search](https://developers.openai.com/api/docs/guides/tools-web-search), [MCP](https://developers.openai.com/api/docs/guides/tools-connectors-mcp), and [computer use](https://developers.openai.com/api/docs/guides/tools-computer-use). This change also lets you manage conversations instead of passing back `previous_response_id`.

15 15 

16### What's changed?16### What's changed?

17 17 

Details

11 11 

12Use this guide to migrate your integration to the [Responses API](https://developers.openai.com/api/docs/guides/migrate-to-responses).12Use this guide to migrate your integration to the [Responses API](https://developers.openai.com/api/docs/guides/migrate-to-responses).

13 13 

14Responses are simpler—send input items and get output items back. With the Responses API, you also get better performance and new features like [deep research](https://developers.openai.com/api/docs/guides/deep-research), [MCP](https://developers.openai.com/api/docs/guides/tools-connectors-mcp), and [computer use](https://developers.openai.com/api/docs/guides/tools-computer-use). This change also lets you manage conversations instead of passing back `previous_response_id`.14Responses are simpler—send input items and get output items back. With the Responses API, you also get better performance and new features like [web search](https://developers.openai.com/api/docs/guides/tools-web-search), [MCP](https://developers.openai.com/api/docs/guides/tools-connectors-mcp), and [computer use](https://developers.openai.com/api/docs/guides/tools-computer-use). This change also lets you manage conversations instead of passing back `previous_response_id`.

15 15 

16### What's changed?16### What's changed?

17 17 

Details

11 11 

12Use this guide to migrate your integration to the [Responses API](https://developers.openai.com/api/docs/guides/migrate-to-responses).12Use this guide to migrate your integration to the [Responses API](https://developers.openai.com/api/docs/guides/migrate-to-responses).

13 13 

14Responses are simpler—send input items and get output items back. With the Responses API, you also get better performance and new features like [deep research](https://developers.openai.com/api/docs/guides/deep-research), [MCP](https://developers.openai.com/api/docs/guides/tools-connectors-mcp), and [computer use](https://developers.openai.com/api/docs/guides/tools-computer-use). This change also lets you manage conversations instead of passing back `previous_response_id`.14Responses are simpler—send input items and get output items back. With the Responses API, you also get better performance and new features like [web search](https://developers.openai.com/api/docs/guides/tools-web-search), [MCP](https://developers.openai.com/api/docs/guides/tools-connectors-mcp), and [computer use](https://developers.openai.com/api/docs/guides/tools-computer-use). This change also lets you manage conversations instead of passing back `previous_response_id`.

15 15 

16### What's changed?16### What's changed?

17 17 

Details

11 11 

12Use this guide to migrate your integration to the [Responses API](https://developers.openai.com/api/docs/guides/migrate-to-responses).12Use this guide to migrate your integration to the [Responses API](https://developers.openai.com/api/docs/guides/migrate-to-responses).

13 13 

14Responses are simpler—send input items and get output items back. With the Responses API, you also get better performance and new features like [deep research](https://developers.openai.com/api/docs/guides/deep-research), [MCP](https://developers.openai.com/api/docs/guides/tools-connectors-mcp), and [computer use](https://developers.openai.com/api/docs/guides/tools-computer-use). This change also lets you manage conversations instead of passing back `previous_response_id`.14Responses are simpler—send input items and get output items back. With the Responses API, you also get better performance and new features like [web search](https://developers.openai.com/api/docs/guides/tools-web-search), [MCP](https://developers.openai.com/api/docs/guides/tools-connectors-mcp), and [computer use](https://developers.openai.com/api/docs/guides/tools-computer-use). This change also lets you manage conversations instead of passing back `previous_response_id`.

15 15 

16### What's changed?16### What's changed?

17 17 

Details

11 11 

12Use this guide to migrate your integration to the [Responses API](https://developers.openai.com/api/docs/guides/migrate-to-responses).12Use this guide to migrate your integration to the [Responses API](https://developers.openai.com/api/docs/guides/migrate-to-responses).

13 13 

14Responses are simpler—send input items and get output items back. With the Responses API, you also get better performance and new features like [deep research](https://developers.openai.com/api/docs/guides/deep-research), [MCP](https://developers.openai.com/api/docs/guides/tools-connectors-mcp), and [computer use](https://developers.openai.com/api/docs/guides/tools-computer-use). This change also lets you manage conversations instead of passing back `previous_response_id`.14Responses are simpler—send input items and get output items back. With the Responses API, you also get better performance and new features like [web search](https://developers.openai.com/api/docs/guides/tools-web-search), [MCP](https://developers.openai.com/api/docs/guides/tools-connectors-mcp), and [computer use](https://developers.openai.com/api/docs/guides/tools-computer-use). This change also lets you manage conversations instead of passing back `previous_response_id`.

15 15 

16### What's changed?16### What's changed?

17 17 

Details

11 11 

12Use this guide to migrate your integration to the [Responses API](https://developers.openai.com/api/docs/guides/migrate-to-responses).12Use this guide to migrate your integration to the [Responses API](https://developers.openai.com/api/docs/guides/migrate-to-responses).

13 13 

14Responses are simpler—send input items and get output items back. With the Responses API, you also get better performance and new features like [deep research](https://developers.openai.com/api/docs/guides/deep-research), [MCP](https://developers.openai.com/api/docs/guides/tools-connectors-mcp), and [computer use](https://developers.openai.com/api/docs/guides/tools-computer-use). This change also lets you manage conversations instead of passing back `previous_response_id`.14Responses are simpler—send input items and get output items back. With the Responses API, you also get better performance and new features like [web search](https://developers.openai.com/api/docs/guides/tools-web-search), [MCP](https://developers.openai.com/api/docs/guides/tools-connectors-mcp), and [computer use](https://developers.openai.com/api/docs/guides/tools-computer-use). This change also lets you manage conversations instead of passing back `previous_response_id`.

15 15 

16### What's changed?16### What's changed?

17 17 

Details

11 11 

12Use this guide to migrate your integration to the [Responses API](https://developers.openai.com/api/docs/guides/migrate-to-responses).12Use this guide to migrate your integration to the [Responses API](https://developers.openai.com/api/docs/guides/migrate-to-responses).

13 13 

14Responses are simpler—send input items and get output items back. With the Responses API, you also get better performance and new features like [deep research](https://developers.openai.com/api/docs/guides/deep-research), [MCP](https://developers.openai.com/api/docs/guides/tools-connectors-mcp), and [computer use](https://developers.openai.com/api/docs/guides/tools-computer-use). This change also lets you manage conversations instead of passing back `previous_response_id`.14Responses are simpler—send input items and get output items back. With the Responses API, you also get better performance and new features like [web search](https://developers.openai.com/api/docs/guides/tools-web-search), [MCP](https://developers.openai.com/api/docs/guides/tools-connectors-mcp), and [computer use](https://developers.openai.com/api/docs/guides/tools-computer-use). This change also lets you manage conversations instead of passing back `previous_response_id`.

15 15 

16### What's changed?16### What's changed?

17 17 

Details

11 11 

12Use this guide to migrate your integration to the [Responses API](https://developers.openai.com/api/docs/guides/migrate-to-responses).12Use this guide to migrate your integration to the [Responses API](https://developers.openai.com/api/docs/guides/migrate-to-responses).

13 13 

14Responses are simpler—send input items and get output items back. With the Responses API, you also get better performance and new features like [deep research](https://developers.openai.com/api/docs/guides/deep-research), [MCP](https://developers.openai.com/api/docs/guides/tools-connectors-mcp), and [computer use](https://developers.openai.com/api/docs/guides/tools-computer-use). This change also lets you manage conversations instead of passing back `previous_response_id`.14Responses are simpler—send input items and get output items back. With the Responses API, you also get better performance and new features like [web search](https://developers.openai.com/api/docs/guides/tools-web-search), [MCP](https://developers.openai.com/api/docs/guides/tools-connectors-mcp), and [computer use](https://developers.openai.com/api/docs/guides/tools-computer-use). This change also lets you manage conversations instead of passing back `previous_response_id`.

15 15 

16### What's changed?16### What's changed?

17 17 

Details

11 11 

12Use this guide to migrate your integration to the [Responses API](https://developers.openai.com/api/docs/guides/migrate-to-responses).12Use this guide to migrate your integration to the [Responses API](https://developers.openai.com/api/docs/guides/migrate-to-responses).

13 13 

14Responses are simpler—send input items and get output items back. With the Responses API, you also get better performance and new features like [deep research](https://developers.openai.com/api/docs/guides/deep-research), [MCP](https://developers.openai.com/api/docs/guides/tools-connectors-mcp), and [computer use](https://developers.openai.com/api/docs/guides/tools-computer-use). This change also lets you manage conversations instead of passing back `previous_response_id`.14Responses are simpler—send input items and get output items back. With the Responses API, you also get better performance and new features like [web search](https://developers.openai.com/api/docs/guides/tools-web-search), [MCP](https://developers.openai.com/api/docs/guides/tools-connectors-mcp), and [computer use](https://developers.openai.com/api/docs/guides/tools-computer-use). This change also lets you manage conversations instead of passing back `previous_response_id`.

15 15 

16### What's changed?16### What's changed?

17 17 

Details

11 11 

12Use this guide to migrate your integration to the [Responses API](https://developers.openai.com/api/docs/guides/migrate-to-responses).12Use this guide to migrate your integration to the [Responses API](https://developers.openai.com/api/docs/guides/migrate-to-responses).

13 13 

14Responses are simpler—send input items and get output items back. With the Responses API, you also get better performance and new features like [deep research](https://developers.openai.com/api/docs/guides/deep-research), [MCP](https://developers.openai.com/api/docs/guides/tools-connectors-mcp), and [computer use](https://developers.openai.com/api/docs/guides/tools-computer-use). This change also lets you manage conversations instead of passing back `previous_response_id`.14Responses are simpler—send input items and get output items back. With the Responses API, you also get better performance and new features like [web search](https://developers.openai.com/api/docs/guides/tools-web-search), [MCP](https://developers.openai.com/api/docs/guides/tools-connectors-mcp), and [computer use](https://developers.openai.com/api/docs/guides/tools-computer-use). This change also lets you manage conversations instead of passing back `previous_response_id`.

15 15 

16### What's changed?16### What's changed?

17 17 

Details

11 11 

12Use this guide to migrate your integration to the [Responses API](https://developers.openai.com/api/docs/guides/migrate-to-responses).12Use this guide to migrate your integration to the [Responses API](https://developers.openai.com/api/docs/guides/migrate-to-responses).

13 13 

14Responses are simpler—send input items and get output items back. With the Responses API, you also get better performance and new features like [deep research](https://developers.openai.com/api/docs/guides/deep-research), [MCP](https://developers.openai.com/api/docs/guides/tools-connectors-mcp), and [computer use](https://developers.openai.com/api/docs/guides/tools-computer-use). This change also lets you manage conversations instead of passing back `previous_response_id`.14Responses are simpler—send input items and get output items back. With the Responses API, you also get better performance and new features like [web search](https://developers.openai.com/api/docs/guides/tools-web-search), [MCP](https://developers.openai.com/api/docs/guides/tools-connectors-mcp), and [computer use](https://developers.openai.com/api/docs/guides/tools-computer-use). This change also lets you manage conversations instead of passing back `previous_response_id`.

15 15 

16### What's changed?16### What's changed?

17 17 

Details

11 11 

12Use this guide to migrate your integration to the [Responses API](https://developers.openai.com/api/docs/guides/migrate-to-responses).12Use this guide to migrate your integration to the [Responses API](https://developers.openai.com/api/docs/guides/migrate-to-responses).

13 13 

14Responses are simpler—send input items and get output items back. With the Responses API, you also get better performance and new features like [deep research](https://developers.openai.com/api/docs/guides/deep-research), [MCP](https://developers.openai.com/api/docs/guides/tools-connectors-mcp), and [computer use](https://developers.openai.com/api/docs/guides/tools-computer-use). This change also lets you manage conversations instead of passing back `previous_response_id`.14Responses are simpler—send input items and get output items back. With the Responses API, you also get better performance and new features like [web search](https://developers.openai.com/api/docs/guides/tools-web-search), [MCP](https://developers.openai.com/api/docs/guides/tools-connectors-mcp), and [computer use](https://developers.openai.com/api/docs/guides/tools-computer-use). This change also lets you manage conversations instead of passing back `previous_response_id`.

15 15 

16### What's changed?16### What's changed?

17 17 

Details

11 11 

12Use this guide to migrate your integration to the [Responses API](https://developers.openai.com/api/docs/guides/migrate-to-responses).12Use this guide to migrate your integration to the [Responses API](https://developers.openai.com/api/docs/guides/migrate-to-responses).

13 13 

14Responses are simpler—send input items and get output items back. With the Responses API, you also get better performance and new features like [deep research](https://developers.openai.com/api/docs/guides/deep-research), [MCP](https://developers.openai.com/api/docs/guides/tools-connectors-mcp), and [computer use](https://developers.openai.com/api/docs/guides/tools-computer-use). This change also lets you manage conversations instead of passing back `previous_response_id`.14Responses are simpler—send input items and get output items back. With the Responses API, you also get better performance and new features like [web search](https://developers.openai.com/api/docs/guides/tools-web-search), [MCP](https://developers.openai.com/api/docs/guides/tools-connectors-mcp), and [computer use](https://developers.openai.com/api/docs/guides/tools-computer-use). This change also lets you manage conversations instead of passing back `previous_response_id`.

15 15 

16### What's changed?16### What's changed?

17 17 

Details

11 11 

12Use this guide to migrate your integration to the [Responses API](https://developers.openai.com/api/docs/guides/migrate-to-responses).12Use this guide to migrate your integration to the [Responses API](https://developers.openai.com/api/docs/guides/migrate-to-responses).

13 13 

14Responses are simpler—send input items and get output items back. With the Responses API, you also get better performance and new features like [deep research](https://developers.openai.com/api/docs/guides/deep-research), [MCP](https://developers.openai.com/api/docs/guides/tools-connectors-mcp), and [computer use](https://developers.openai.com/api/docs/guides/tools-computer-use). This change also lets you manage conversations instead of passing back `previous_response_id`.14Responses are simpler—send input items and get output items back. With the Responses API, you also get better performance and new features like [web search](https://developers.openai.com/api/docs/guides/tools-web-search), [MCP](https://developers.openai.com/api/docs/guides/tools-connectors-mcp), and [computer use](https://developers.openai.com/api/docs/guides/tools-computer-use). This change also lets you manage conversations instead of passing back `previous_response_id`.

15 15 

16### What's changed?16### What's changed?

17 17 

Details

11 11 

12Use this guide to migrate your integration to the [Responses API](https://developers.openai.com/api/docs/guides/migrate-to-responses).12Use this guide to migrate your integration to the [Responses API](https://developers.openai.com/api/docs/guides/migrate-to-responses).

13 13 

14Responses are simpler—send input items and get output items back. With the Responses API, you also get better performance and new features like [deep research](https://developers.openai.com/api/docs/guides/deep-research), [MCP](https://developers.openai.com/api/docs/guides/tools-connectors-mcp), and [computer use](https://developers.openai.com/api/docs/guides/tools-computer-use). This change also lets you manage conversations instead of passing back `previous_response_id`.14Responses are simpler—send input items and get output items back. With the Responses API, you also get better performance and new features like [web search](https://developers.openai.com/api/docs/guides/tools-web-search), [MCP](https://developers.openai.com/api/docs/guides/tools-connectors-mcp), and [computer use](https://developers.openai.com/api/docs/guides/tools-computer-use). This change also lets you manage conversations instead of passing back `previous_response_id`.

15 15 

16### What's changed?16### What's changed?

17 17 

Details

11 11 

12Use this guide to migrate your integration to the [Responses API](https://developers.openai.com/api/docs/guides/migrate-to-responses).12Use this guide to migrate your integration to the [Responses API](https://developers.openai.com/api/docs/guides/migrate-to-responses).

13 13 

14Responses are simpler—send input items and get output items back. With the Responses API, you also get better performance and new features like [deep research](https://developers.openai.com/api/docs/guides/deep-research), [MCP](https://developers.openai.com/api/docs/guides/tools-connectors-mcp), and [computer use](https://developers.openai.com/api/docs/guides/tools-computer-use). This change also lets you manage conversations instead of passing back `previous_response_id`.14Responses are simpler—send input items and get output items back. With the Responses API, you also get better performance and new features like [web search](https://developers.openai.com/api/docs/guides/tools-web-search), [MCP](https://developers.openai.com/api/docs/guides/tools-connectors-mcp), and [computer use](https://developers.openai.com/api/docs/guides/tools-computer-use). This change also lets you manage conversations instead of passing back `previous_response_id`.

15 15 

16### What's changed?16### What's changed?

17 17 

Details

11 11 

12Use this guide to migrate your integration to the [Responses API](https://developers.openai.com/api/docs/guides/migrate-to-responses).12Use this guide to migrate your integration to the [Responses API](https://developers.openai.com/api/docs/guides/migrate-to-responses).

13 13 

14Responses are simpler—send input items and get output items back. With the Responses API, you also get better performance and new features like [deep research](https://developers.openai.com/api/docs/guides/deep-research), [MCP](https://developers.openai.com/api/docs/guides/tools-connectors-mcp), and [computer use](https://developers.openai.com/api/docs/guides/tools-computer-use). This change also lets you manage conversations instead of passing back `previous_response_id`.14Responses are simpler—send input items and get output items back. With the Responses API, you also get better performance and new features like [web search](https://developers.openai.com/api/docs/guides/tools-web-search), [MCP](https://developers.openai.com/api/docs/guides/tools-connectors-mcp), and [computer use](https://developers.openai.com/api/docs/guides/tools-computer-use). This change also lets you manage conversations instead of passing back `previous_response_id`.

15 15 

16### What's changed?16### What's changed?

17 17 

Details

172 172 

173 The list of completion choices the model generated for the input prompt.173 The list of completion choices the model generated for the input prompt.

174 174 

175 - `finish_reason: "stop" or "length" or "content_filter"`175 - `finish_reason: "stop" or "length" or "content_filter" or null`

176 176 

177 The reason the model stopped generating tokens. This will be `stop` if the model hit a natural stop point or a provided stop sequence,177 The reason the model stopped generating tokens. This will be `stop` if the model hit a natural stop point or a provided stop sequence,

178 `length` if the maximum number of tokens specified in the request was reached,178 `length` if the maximum number of tokens specified in the request was reached,

179 or `content_filter` if content was omitted due to a flag from our content filters.179 or `content_filter` if content was omitted due to a flag from our content filters. The value is null while a streamed completion is unfinished.

180 180 

181 - `"stop"`181 - `"stop"`

182 182 

Details

2078 2078 

2079 An alternative to temperature for nucleus sampling; 1.0 includes all tokens.2079 An alternative to temperature for nucleus sampling; 1.0 includes all tokens.

2080 2080 

2081- `error: EvalAPIError`2081- `error: EvalAPIError or null`

2082 2082 

2083 An object representing an error response from the Eval API.2083 An object representing an error response from the Eval API.

2084 2084 


2103 Keys are strings with a maximum length of 64 characters. Values are strings2103 Keys are strings with a maximum length of 64 characters. Values are strings

2104 with a maximum length of 512 characters.2104 with a maximum length of 512 characters.

2105 2105 

2106- `model: string`2106- `model: string or null`

2107 2107 

2108 The model that is evaluated, if applicable.2108 The model that is evaluated, if applicable.

2109 2109 

2110- `name: string`2110- `name: string or null`

2111 2111 

2112 The name of the evaluation run.2112 The name of the evaluation run.

2113 2113 


2117 2117 

2118 - `"eval.run"`2118 - `"eval.run"`

2119 2119 

2120- `per_model_usage: array of object { cached_tokens, completion_tokens, invocation_count, 3 more }`2120- `per_model_usage: array of object { cached_tokens, completion_tokens, invocation_count, 3 more } or null`

2121 2121 

2122 Usage statistics for each model during the evaluation run.2122 Usage statistics for each model during the evaluation run.

2123 2123 


2145 2145 

2146 The total number of tokens used.2146 The total number of tokens used.

2147 2147 

2148- `per_testing_criteria_results: array of object { failed, passed, testing_criteria }`2148- `per_testing_criteria_results: array of object { failed, passed, testing_criteria } or null`

2149 2149 

2150 Results per testing criteria applied during the evaluation run.2150 Results per testing criteria applied during the evaluation run.

2151 2151 

Details

611 611 

612 A formula to calculate the output based on grader results.612 A formula to calculate the output based on grader results.

613 613 

614 - `graders: StringCheckGrader or TextSimilarityGrader or PythonGrader or 2 more`614 - `graders: map[StringCheckGrader or TextSimilarityGrader or PythonGrader or 2 more]`

615 

616 A StringCheckGrader object that performs a string comparison between input and reference using a specified operation.

617 615 

618 - `StringCheckGrader object { input, name, operation, 2 more }`616 - `StringCheckGrader object { input, name, operation, 2 more }`

619 617 

Details

296 296 

297 A formula to calculate the output based on grader results.297 A formula to calculate the output based on grader results.

298 298 

299 - `graders: StringCheckGrader or TextSimilarityGrader or PythonGrader or 2 more`299 - `graders: map[StringCheckGrader or TextSimilarityGrader or PythonGrader or 2 more]`

300 

301 A StringCheckGrader object that performs a string comparison between input and reference using a specified operation.

302 300 

303 - `StringCheckGrader object { input, name, operation, 2 more }`301 - `StringCheckGrader object { input, name, operation, 2 more }`

304 302 

Details

4479 4479 

4480 - `string`4480 - `string`

4481 4481 

4482 - `"gpt-6-astra" or "gpt-6-sol" or "gpt-6-luna" or 85 more`4482 - `"gpt-6-astra" or "gpt-6.1-sol" or "gpt-6-sol" or 86 more`

4483 4483 

4484 - `"gpt-6-astra"`4484 - `"gpt-6-astra"`

4485 4485 

4486 - `"gpt-6.1-sol"`

4487 

4486 - `"gpt-6-sol"`4488 - `"gpt-6-sol"`

4487 4489 

4488 - `"gpt-6-luna"`4490 - `"gpt-6-luna"`

Details

10 10 

11### Body Parameters11### Body Parameters

12 12 

13- `model: "gpt-6-astra" or "gpt-6-sol" or "gpt-6-luna" or 105 more or string or null`13- `model: "gpt-6-astra" or "gpt-6.1-sol" or "gpt-6-sol" or 106 more or string or null`

14 14 

15 Model ID used to generate the response, like `gpt-6-astra`. OpenAI offers a wide range of models with different capabilities, performance characteristics, and price points. Refer to the [model guide](/api/docs/models) to browse and compare available models.15 Model ID used to generate the response, like `gpt-6-astra`. OpenAI offers a wide range of models with different capabilities, performance characteristics, and price points. Refer to the [model guide](/api/docs/models) to browse and compare available models.

16 16 

17 - `"gpt-6-astra" or "gpt-6-sol" or "gpt-6-luna" or 105 more`17 - `"gpt-6-astra" or "gpt-6.1-sol" or "gpt-6-sol" or 106 more`

18 18 

19 Model ID used to generate the response, like `gpt-6-astra`. OpenAI offers a wide range of models with different capabilities, performance characteristics, and price points. Refer to the [model guide](/api/docs/models) to browse and compare available models.19 Model ID used to generate the response, like `gpt-6-astra`. OpenAI offers a wide range of models with different capabilities, performance characteristics, and price points. Refer to the [model guide](/api/docs/models) to browse and compare available models.

20 20 

21 - `"gpt-6-astra"`21 - `"gpt-6-astra"`

22 22 

23 - `"gpt-6.1-sol"`

24 

23 - `"gpt-6-sol"`25 - `"gpt-6-sol"`

24 26 

25 - `"gpt-6-luna"`27 - `"gpt-6-luna"`

Details

4448 4448 

4449 - `string`4449 - `string`

4450 4450 

4451 - `"gpt-6-astra" or "gpt-6-sol" or "gpt-6-luna" or 85 more`4451 - `"gpt-6-astra" or "gpt-6.1-sol" or "gpt-6-sol" or 86 more`

4452 4452 

4453 - `"gpt-6-astra"`4453 - `"gpt-6-astra"`

4454 4454 

4455 - `"gpt-6.1-sol"`

4456 

4455 - `"gpt-6-sol"`4457 - `"gpt-6-sol"`

4456 4458 

4457 - `"gpt-6-luna"`4459 - `"gpt-6-luna"`


10587 10589 

10588 - `string`10590 - `string`

10589 10591 

10590 - `"gpt-6-astra" or "gpt-6-sol" or "gpt-6-luna" or 85 more`10592 - `"gpt-6-astra" or "gpt-6.1-sol" or "gpt-6-sol" or 86 more`

10591 10593 

10592 - `"gpt-6-astra"`10594 - `"gpt-6-astra"`

10593 10595 

10596 - `"gpt-6.1-sol"`

10597 

10594 - `"gpt-6-sol"`10598 - `"gpt-6-sol"`

10595 10599 

10596 - `"gpt-6-luna"`10600 - `"gpt-6-luna"`

Details

4523 4523 

4524 - `string`4524 - `string`

4525 4525 

4526 - `"gpt-6-astra" or "gpt-6-sol" or "gpt-6-luna" or 85 more`4526 - `"gpt-6-astra" or "gpt-6.1-sol" or "gpt-6-sol" or 86 more`

4527 4527 

4528 - `"gpt-6-astra"`4528 - `"gpt-6-astra"`

4529 4529 

4530 - `"gpt-6.1-sol"`

4531 

4530 - `"gpt-6-sol"`4532 - `"gpt-6-sol"`

4531 4533 

4532 - `"gpt-6-luna"`4534 - `"gpt-6-luna"`