Calls

Accept call

$ openai realtime:calls accept

post /realtime/calls/{call_id}/accept

Accept an incoming SIP call and configure the realtime session that will handle it.

Parameters

--call-id: string

The identifier for the call provided in the realtime.call.incoming webhook.
--type: "realtime"

The type of session to create. Always realtime for the Realtime API.
--audio: optional object { input, output }

Configuration for input and output audio.
--include: optional array of "item.input_audio_transcription.logprobs"

Additional fields to include in server outputs.

item.input_audio_transcription.logprobs: Include logprobs for input audio transcription.
--instructions: optional string

The default system instructions (i.e. system message) prepended to model calls. This field allows the client to guide the model on desired responses. The model can be instructed on response content and format, (e.g. "be extremely succinct", "act friendly", "here are examples of good responses") and on audio behavior (e.g. "talk quickly", "inject emotion into your voice", "laugh frequently"). The instructions are not guaranteed to be followed by the model, but they provide guidance to the model on the desired behavior.

Note that the server sets default instructions which will be used if this field is not set and are visible in the session.created event at the start of the session.
--max-output-tokens: optional number or "inf"

Maximum number of output tokens for a single assistant response, inclusive of tool calls. Provide an integer between 1 and 4096 to limit output tokens, or inf for the maximum available tokens for a given model. Defaults to inf.
--model: optional string or "gpt-realtime" or "gpt-realtime-1.5" or "gpt-realtime-2" or 14 more

The Realtime model used for this session.
--output-modality: optional array of "text" or "audio"

The set of modalities the model can respond with. It defaults to ["audio"], indicating that the model will respond with audio plus a transcript. ["text"] can be used to make the model respond with text only. It is not possible to request both text and audio at the same time.
--parallel-tool-calls: optional boolean

Whether the model may call multiple tools in parallel. Only supported by reasoning Realtime models such as gpt-realtime-2.
--prompt: optional object { id, variables, version }

Reference to a prompt template and its variables. Learn more.
--reasoning: optional object { effort }

Configuration for reasoning-capable Realtime models such as gpt-realtime-2.
--tool-choice: optional ToolChoiceOptions or ToolChoiceFunction or ToolChoiceMcp

How the model chooses tools. Provide one of the string modes or force a specific function/MCP tool.
--tool: optional array of RealtimeToolsConfigUnion

Tools available to the model.
--tracing: optional "auto" or object { group_id, metadata, workflow_name }

Realtime API can write session traces to the Traces Dashboard. Set to null to disable tracing. Once tracing is enabled for a session, the configuration cannot be modified.

auto will create a trace for the session with default values for the workflow name, group id, and metadata.
--truncation: optional "auto" or "disabled" or RealtimeTruncationRetentionRatio

When the number of tokens in a conversation exceeds the model's input token limit, the conversation be truncated, meaning messages (starting from the oldest) will not be included in the model's context. A 32k context model with 4,096 max output tokens can only include 28,224 tokens in the context before truncation occurs.

Clients can configure truncation behavior to truncate with a lower max token limit, which is an effective way to control token usage and cost.

Truncation will reduce the number of cached tokens on the next turn (busting the cache), since messages are dropped from the beginning of the context. However, clients can also configure truncation to retain messages up to a fraction of the maximum context size, which will reduce the need for future truncations and thus improve the cache rate.

Truncation can be disabled entirely, which means the server will never truncate but would instead return an error if the conversation exceeds the model's input token limit.

Example

openai realtime:calls accept \
  --api-key 'My API Key' \
  --call-id call_id \
  --type realtime

Hang up call

$ openai realtime:calls hangup

post /realtime/calls/{call_id}/hangup

End an active Realtime API call, whether it was initiated over SIP or WebRTC.

Parameters

--call-id: string

The identifier for the call. For SIP calls, use the value provided in the realtime.call.incoming webhook. For WebRTC sessions, reuse the call ID returned in the Location header when creating the call with POST /v1/realtime/calls.

Example

openai realtime:calls hangup \
  --api-key 'My API Key' \
  --call-id call_id

Refer call

$ openai realtime:calls refer

post /realtime/calls/{call_id}/refer

Transfer an active SIP call to a new destination using the SIP REFER verb.

Parameters

--call-id: string

The identifier for the call provided in the realtime.call.incoming webhook.
--target-uri: string

URI that should appear in the SIP Refer-To header. Supports values like tel:+14155550123 or sip:agent@example.com.

Example

openai realtime:calls refer \
  --api-key 'My API Key' \
  --call-id call_id \
  --target-uri tel:+14155550123

Reject call

$ openai realtime:calls reject

post /realtime/calls/{call_id}/reject

Decline an incoming SIP call by returning a SIP status code to the caller.

Parameters

--call-id: string

The identifier for the call provided in the realtime.call.incoming webhook.
--status-code: optional number

SIP response code to send back to the caller. Defaults to 603 (Decline) when omitted.

Example

openai realtime:calls reject \
  --api-key 'My API Key' \
  --call-id call_id

cli/resources/realtime/subresources/calls/index.md +191 −0 created

1# Calls

3## Accept call

5`$ openai realtime:calls accept`

7**post** `/realtime/calls/{call_id}/accept`

9Accept an incoming SIP call and configure the realtime session that will

10handle it.

12### Parameters

14- `--call-id: string`

16 The identifier for the call provided in the

17 [`realtime.call.incoming`](https://platform.openai.com/docs/api-reference/webhook-events/realtime/call/incoming)

18 webhook.

20- `--type: "realtime"`

22 The type of session to create. Always `realtime` for the Realtime API.

24- `--audio: optional object { input, output }`

26 Configuration for input and output audio.

28- `--include: optional array of "item.input_audio_transcription.logprobs"`

30 Additional fields to include in server outputs.

32 `item.input_audio_transcription.logprobs`: Include logprobs for input audio transcription.

34- `--instructions: optional string`

36 The default system instructions (i.e. system message) prepended to model calls. This field allows the client to guide the model on desired responses. The model can be instructed on response content and format, (e.g. "be extremely succinct", "act friendly", "here are examples of good responses") and on audio behavior (e.g. "talk quickly", "inject emotion into your voice", "laugh frequently"). The instructions are not guaranteed to be followed by the model, but they provide guidance to the model on the desired behavior.

38 Note that the server sets default instructions which will be used if this field is not set and are visible in the `session.created` event at the start of the session.

40- `--max-output-tokens: optional number or "inf"`

42 Maximum number of output tokens for a single assistant response,

43 inclusive of tool calls. Provide an integer between 1 and 4096 to

44 limit output tokens, or `inf` for the maximum available tokens for a

45 given model. Defaults to `inf`.

47- `--model: optional string or "gpt-realtime" or "gpt-realtime-1.5" or "gpt-realtime-2" or 14 more`

49 The Realtime model used for this session.

51- `--output-modality: optional array of "text" or "audio"`

53 The set of modalities the model can respond with. It defaults to `["audio"]`, indicating

54 that the model will respond with audio plus a transcript. `["text"]` can be used to make

55 the model respond with text only. It is not possible to request both `text` and `audio` at the same time.

57- `--parallel-tool-calls: optional boolean`

59 Whether the model may call multiple tools in parallel. Only supported by

60 reasoning Realtime models such as `gpt-realtime-2`.

62- `--prompt: optional object { id, variables, version }`

64 Reference to a prompt template and its variables.

65 [Learn more](https://platform.openai.com/docs/guides/text?api-mode=responses#reusable-prompts).

67- `--reasoning: optional object { effort }`

69 Configuration for reasoning-capable Realtime models such as `gpt-realtime-2`.

71- `--tool-choice: optional ToolChoiceOptions or ToolChoiceFunction or ToolChoiceMcp`

73 How the model chooses tools. Provide one of the string modes or force a specific

74 function/MCP tool.

76- `--tool: optional array of RealtimeToolsConfigUnion`

78 Tools available to the model.

80- `--tracing: optional "auto" or object { group_id, metadata, workflow_name }`

82 Realtime API can write session traces to the [Traces Dashboard](https://platform.openai.com/logs?api=traces). Set to null to disable tracing. Once

83 tracing is enabled for a session, the configuration cannot be modified.

85 `auto` will create a trace for the session with default values for the

86 workflow name, group id, and metadata.

88- `--truncation: optional "auto" or "disabled" or RealtimeTruncationRetentionRatio`

90 When the number of tokens in a conversation exceeds the model's input token limit, the conversation be truncated, meaning messages (starting from the oldest) will not be included in the model's context. A 32k context model with 4,096 max output tokens can only include 28,224 tokens in the context before truncation occurs.

92 Clients can configure truncation behavior to truncate with a lower max token limit, which is an effective way to control token usage and cost.

94 Truncation will reduce the number of cached tokens on the next turn (busting the cache), since messages are dropped from the beginning of the context. However, clients can also configure truncation to retain messages up to a fraction of the maximum context size, which will reduce the need for future truncations and thus improve the cache rate.

96 Truncation can be disabled entirely, which means the server will never truncate but would instead return an error if the conversation exceeds the model's input token limit.

98### Example

100```cli

101openai realtime:calls accept \

102 --api-key 'My API Key' \

103 --call-id call_id \

104 --type realtime

105```

106

107## Hang up call

108

109`$ openai realtime:calls hangup`

110

111**post** `/realtime/calls/{call_id}/hangup`

112

113End an active Realtime API call, whether it was initiated over SIP or

114WebRTC.

115

116### Parameters

117

118- `--call-id: string`

119

120 The identifier for the call. For SIP calls, use the value provided in the

121 [`realtime.call.incoming`](https://platform.openai.com/docs/api-reference/webhook-events/realtime/call/incoming)

122 webhook. For WebRTC sessions, reuse the call ID returned in the `Location`

123 header when creating the call with

124 [`POST /v1/realtime/calls`](https://platform.openai.com/docs/api-reference/realtime/create-call).

125

126### Example

127

128```cli

129openai realtime:calls hangup \

130 --api-key 'My API Key' \

131 --call-id call_id

132```

133

134## Refer call

135

136`$ openai realtime:calls refer`

137

138**post** `/realtime/calls/{call_id}/refer`

139

140Transfer an active SIP call to a new destination using the SIP REFER verb.

141

142### Parameters

143

144- `--call-id: string`

145

146 The identifier for the call provided in the

147 [`realtime.call.incoming`](https://platform.openai.com/docs/api-reference/webhook-events/realtime/call/incoming)

148 webhook.

149

150- `--target-uri: string`

151

152 URI that should appear in the SIP Refer-To header. Supports values like

153 `tel:+14155550123` or `sip:agent@example.com`.

154

155### Example

156

157```cli

158openai realtime:calls refer \

159 --api-key 'My API Key' \

160 --call-id call_id \

161 --target-uri tel:+14155550123

162```

163

164## Reject call

165

166`$ openai realtime:calls reject`

167

168**post** `/realtime/calls/{call_id}/reject`

169

170Decline an incoming SIP call by returning a SIP status code to the caller.

171

172### Parameters

173

174- `--call-id: string`

175

176 The identifier for the call provided in the

177 [`realtime.call.incoming`](https://platform.openai.com/docs/api-reference/webhook-events/realtime/call/incoming)

178 webhook.

179

180- `--status-code: optional number`

181

182 SIP response code to send back to the caller. Defaults to `603` (Decline)

183 when omitted.

184

185### Example

186

187```cli

188openai realtime:calls reject \

189 --api-key 'My API Key' \

190 --call-id call_id

191```

cli/resources/realtime/subresources/calls/index.md 2026-05-05 23:00 UTC to 2026-05-07 21:57 UTC

Calls

Accept call

Parameters

Example

Hang up call

Parameters

Example

Refer call

Parameters

Example

Reject call

Parameters

Example