openapi: 3.0.3 info: title: Retell SDK Add Community Voice Create Web Call API version: 3.0.0 contact: name: Retell Support url: https://www.retellai.com/ email: support@retellai.com license: name: Apache 2.0 url: https://www.apache.org/licenses/LICENSE-2.0.html servers: - url: https://api.retellai.com description: The production server. security: - api_key: [] tags: - name: Create Web Call paths: /v2/create-web-call: post: description: Create a new web call operationId: createWebCall requestBody: required: true content: application/json: schema: type: object required: - agent_id properties: agent_id: type: string minLength: 1 example: oBeDLoLOeuAbiuaMFXRtDOLriTJ5tSxD description: Unique id of agent used for the call. Your agent would contain the LLM Websocket url used for this call. agent_version: $ref: '#/components/schemas/AgentVersionReference' description: The version of the agent to use for the call. agent_override: $ref: '#/components/schemas/AgentOverrideRequest' description: For this particular call, override agent configuration with these settings. This allows you to customize agent behavior for individual calls without modifying the base agent. metadata: type: object description: An arbitrary object for storage purpose only. You can put anything here like your internal customer id associated with the call. Not used for processing. You can later get this field from the call object. retell_llm_dynamic_variables: type: object additionalProperties: {} example: customer_name: John Doe description: Add optional dynamic variables in key value pairs of string that injects into your Response Engine prompt and tool description. Only applicable for Response Engine. current_node_id: type: string example: collect_info description: Start the call at this conversation flow node (stage). Must be a valid node id in the agent's conversation flow. Only applicable when the agent uses conversation flow as the response engine. Ignored for retell-llm agents. nullable: true current_state: type: string example: information_collection description: Start the conversation in this state (stage). Must be a valid state name in the agent's Retell LLM. Only applicable when the agent uses Retell LLM with states. Ignored for conversation-flow agents. nullable: true responses: '201': description: Successfully created a web call. content: application/json: schema: $ref: '#/components/schemas/V2WebCallResponse' '400': $ref: '#/components/responses/BadRequest' '401': $ref: '#/components/responses/Unauthorized' '402': $ref: '#/components/responses/PaymentRequired' '422': $ref: '#/components/responses/UnprocessableContent' '429': $ref: '#/components/responses/TooManyRequests' '500': $ref: '#/components/responses/InternalServerError' tags: - Create Web Call components: schemas: CallLatency: type: object properties: p50: type: number description: 50 percentile of latency, measured in milliseconds. example: 800 p90: type: number description: 90 percentile of latency, measured in milliseconds. example: 1200 p95: type: number description: 95 percentile of latency, measured in milliseconds. example: 1500 p99: type: number description: 99 percentile of latency, measured in milliseconds. example: 2500 max: type: number description: Maximum latency in the call, measured in milliseconds. example: 2700 min: type: number description: Minimum latency in the call, measured in milliseconds. example: 500 num: type: number description: Number of data points (number of times latency is tracked). example: 10 values: type: array items: type: number description: All the latency data points in the call, measured in milliseconds. ToolCallResultUtterance: type: object required: - role - tool_call_id - content properties: role: type: string enum: - tool_call_result description: This is the result of a tool call. tool_call_id: type: string description: Tool call id, globally unique. content: type: string description: Result of the tool call, can be a string, a stringified json, etc. successful: type: boolean description: Whether the tool call was successful. CallPresetAnalysisData: type: object required: - type - name description: System preset for post-call analysis (voice agents). Use in post_call_analysis_data to override prompts or mark fields optional. properties: type: type: string enum: - system-presets description: Identifies this item as a system preset. name: type: string enum: - call_summary - call_successful - user_sentiment description: Preset identifier for voice agent analysis. example: call_summary description: type: string minLength: 1 description: Prompt or description for this preset. required: type: boolean description: If false, this field is optional in the analysis. If true or unset, the field is required. conditional_prompt: type: string description: Optional instruction to help decide whether this field needs to be populated. If not set, the field is always included. CallAnalysis: type: object properties: call_summary: type: string example: The agent called the user to ask question about his purchase inquiry. The agent asked several questions regarding his preference and asked if user would like to book an appointment. The user happily agreed and scheduled an appointment next Monday 10am. description: A high level summary of the call. in_voicemail: type: boolean example: false description: Whether the call is entered voicemail. user_sentiment: type: string enum: - Negative - Positive - Neutral - Unknown example: Positive description: Sentiment of the user in the call. call_successful: type: boolean example: true description: Whether the agent seems to have a successful call with the user, where the agent finishes the task, and the call was complete without being cutoff. custom_analysis_data: type: object description: Custom analysis data that was extracted based on the schema defined in agent post call analysis data. Can be empty if nothing is specified. VoicemailActionBridgeTransfer: type: object properties: type: type: string enum: - bridge_transfer example: bridge_transfer required: - type UtteranceOrToolCall: oneOf: - $ref: '#/components/schemas/Utterance' - $ref: '#/components/schemas/ToolCallInvocationUtterance' - $ref: '#/components/schemas/ToolCallResultUtterance' - $ref: '#/components/schemas/NodeTransitionUtterance' - $ref: '#/components/schemas/DTMFUtterance' VoicemailActionPrompt: type: object properties: type: type: string enum: - prompt example: prompt text: type: string example: Summarize the call in 2 sentences. description: The prompt used to generate the text to be spoken when the call is detected to be in voicemail. required: - type - text NodeTransitionUtterance: type: object required: - role - former_node_id - former_node_name - new_node_id - new_node_name properties: role: type: string enum: - node_transition description: This is result of a node transition former_node_id: type: string description: Former node id former_node_name: type: string description: Former node name new_node_id: type: string description: New node id new_node_name: type: string description: New node name transition_type: type: string enum: - global - global_go_back - interrupt_go_back - normal description: How this node was reached. "global" means a global node transition, "global_go_back" means returning from a global node, "interrupt_go_back" means going back due to user interruption, and "normal" means a regular edge transition. LanguageLegacy: oneOf: - $ref: '#/components/schemas/Language' - type: string enum: - multi example: en-US description: Legacy single-string language format. Accepts any concrete locale from `Language`, plus the special scalar value `multi` for multilingual support. If unset, will use default value `en-US`. Language: type: string example: en-US enum: - en-US - en-IN - en-GB - en-AU - en-NZ - de-DE - es-ES - es-419 - hi-IN - fr-FR - fr-CA - ja-JP - pt-PT - pt-BR - zh-CN - ru-RU - it-IT - ko-KR - nl-NL - nl-BE - pl-PL - tr-TR - vi-VN - ro-RO - bg-BG - ca-ES - th-TH - da-DK - fi-FI - el-GR - hu-HU - id-ID - no-NO - sk-SK - sv-SE - lt-LT - lv-LV - cs-CZ - ms-MY - af-ZA - ar-SA - az-AZ - bs-BA - cy-GB - fa-IR - fil-PH - gl-ES - he-IL - hr-HR - hy-AM - is-IS - kk-KZ - kn-IN - mk-MK - mr-IN - ne-NP - sl-SI - sr-RS - sw-KE - ta-IN - ur-IN - yue-CN - uk-UA description: Specifies what language (and dialect) the agent will operate in. For instance, selecting `en-GB` optimizes speech recognition for British English and indexes knowledge bases with English. If unset, will use default value `en-US`. This enum does not include the legacy scalar value `multi`. ConversationFlowOverride: type: object description: Override properties for conversation flow configuration in agent override requests. properties: model_choice: $ref: '#/components/schemas/ModelChoice' description: The model choice for the conversation flow. model_temperature: type: number minimum: 0 maximum: 1 example: 0.7 description: Controls the randomness of the model's responses. Lower values make responses more deterministic. nullable: true tool_call_strict_mode: type: boolean example: true description: Whether to use strict mode for tool calls. Only applicable when using certain supported models. nullable: true knowledge_base_ids: type: array items: type: string example: - kb_001 - kb_002 description: Knowledge base IDs for RAG (Retrieval-Augmented Generation). nullable: true kb_config: type: object $ref: '#/components/schemas/KBConfig' description: Knowledge base configuration for RAG retrieval. start_speaker: type: string enum: - user - agent example: agent description: Who starts the conversation - user or agent. begin_after_user_silence_ms: type: integer example: 2000 description: If set, the AI will begin the conversation after waiting for the user for the duration (in milliseconds) specified by this attribute. This only applies if the agent is configured to wait for the user to speak first. If not set, the agent will wait indefinitely for the user to speak. nullable: true ModelChoice: oneOf: - $ref: '#/components/schemas/ModelChoiceCascading' ProductCost: type: object required: - product - cost properties: product: type: string description: Product name that has a cost associated with it. example: elevenlabs_tts unit_price: type: number description: Unit price of the product in cents per second. example: 1 cost: type: number description: Cost for the product in cents for the duration of the call. example: 60 is_transfer_leg_cost: type: boolean description: True if this cost item is for a transfer segment. AgentRequest: type: object properties: response_engine: $ref: '#/components/schemas/ResponseEngine' description: The Response Engine to attach to the agent. It is used to generate responses for the agent. You need to create a Response Engine first before attaching it to an agent. example: type: retell-llm llm_id: llm_234sdertfsdsfsdf version: 0 agent_name: type: string example: Jarvis description: The name of the agent. Only used for your own reference. nullable: true version_description: type: string example: Customer support agent for handling product inquiries description: Optional description of the agent version. Used for your own reference and documentation. nullable: true voice_id: type: string example: retell-Cimo description: Unique voice id used for the agent. Find list of available voices and their preview in Dashboard. voice_model: type: string enum: - eleven_turbo_v2 - eleven_flash_v2 - eleven_turbo_v2_5 - eleven_flash_v2_5 - eleven_multilingual_v2 - eleven_v3 - sonic-3 - sonic-3-latest - tts-1 - gpt-4o-mini-tts - speech-02-turbo - speech-2.8-turbo - s1 - s2-pro - null description: Select the voice model used for the selected voice. Each provider has a set of available voice models. Set to null to remove voice model selection, and default ones will apply. Check out dashboard for more details of each voice model. nullable: true fallback_voice_ids: type: array items: type: string example: - cartesia-Cimo - minimax-Cimo description: When TTS provider for the selected voice is experiencing outages, we would use fallback voices listed here for the agent. Voice id and the fallback voice ids must be from different TTS providers. The system would go through the list in order, if the first one in the list is also having outage, it would use the next one. Set to null to remove voice fallback for the agent. nullable: true voice_temperature: type: number example: 1 description: Controls how stable the voice is. Value ranging from [0,2]. Lower value means more stable, and higher value means more variant speech generation. Check the dashboard to see what provider supports this feature. If unset, default value 1 will apply. voice_speed: type: number minimum: 0.5 maximum: 2 example: 1 description: Controls speed of voice. Value ranging from [0.5,2]. Lower value means slower speech, while higher value means faster speech rate. If unset, default value 1 will apply. enable_dynamic_voice_speed: type: boolean example: true description: If set to true, will enable dynamic voice speed adjustment based on the user's speech rate and conversation context. If unset, default value false will apply. enable_dynamic_responsiveness: type: boolean example: true description: If set to true, the agent will dynamically adjust how quickly it responds based on the user's speech rate and past turn-taking behavior in the call. If unset, default value false will apply. volume: type: number example: 1 description: If set, will control the volume of the agent. Value ranging from [0,2]. Lower value means quieter agent speech, while higher value means louder agent speech. If unset, default value 1 will apply. voice_emotion: type: string nullable: true enum: - calm - sympathetic - happy - sad - angry - fearful - surprised - null example: calm description: 'Controls the emotional tone of the agent''s voice. Currently supported for Cartesia and Minimax TTS providers. If unset, no emotion will be used. ' responsiveness: type: number minimum: 0 maximum: 1 example: 1 description: Controls how responsive is the agent. Value ranging from [0,1]. Lower value means less responsive agent (wait more, respond slower), while higher value means faster exchanges (respond when it can). If unset, default value 1 will apply. interruption_sensitivity: type: number minimum: 0 maximum: 1 example: 1 description: Controls how sensitive the agent is to user interruptions. Value ranging from [0,1]. Lower value means it will take longer / more words for user to interrupt agent, while higher value means it's easier for user to interrupt agent. If unset, default value 1 will apply. When this is set to 0, agent would never be interrupted. enable_backchannel: type: boolean example: true description: Controls whether the agent would backchannel (agent interjects the speaker with phrases like "yeah", "uh-huh" to signify interest and engagement). Backchannel when enabled tends to show up more in longer user utterances. If not set, agent will not backchannel. backchannel_frequency: type: number example: 0.9 description: Only applicable when enable_backchannel is true. Controls how often the agent would backchannel when a backchannel is possible. Value ranging from [0,1]. Lower value means less frequent backchannel, while higher value means more frequent backchannel. If unset, default value 0.8 will apply. backchannel_words: type: array items: type: string example: - yeah - uh-huh description: Only applicable when enable_backchannel is true. A list of words that the agent would use as backchannel. If not set, default backchannel words will apply. Check out [backchannel default words](/agent/interaction-configuration#backchannel) for more details. Note that certain voices do not work too well with certain words, so it's recommended to experiment before adding any words. nullable: true reminder_trigger_ms: type: number example: 10000 description: If set (in milliseconds), will trigger a reminder to the agent to speak if the user has been silent for the specified duration after some agent speech. Must be a positive number. If unset, default value of 10000 ms (10 s) will apply. reminder_max_count: type: integer example: 2 description: If set, controls how many times agent would remind user when user is unresponsive. Must be a non negative integer. If unset, default value of 1 will apply (remind once). Set to 0 to disable agent from reminding. ambient_sound: type: string enum: - coffee-shop - convention-hall - summer-outdoor - mountain-outdoor - static-noise - call-center - null description: 'If set, will add ambient environment sound to the call to make experience more realistic. Currently supports the following options: - `coffee-shop`: Coffee shop ambience with people chatting in background. [Listen to Ambience](https://retell-utils-public.s3.us-west-2.amazonaws.com/coffee-shop.wav) - `convention-hall`: Convention hall ambience, with some echo and people chatting in background. [Listen to Ambience](https://retell-utils-public.s3.us-west-2.amazonaws.com/convention-hall.wav) - `summer-outdoor`: Summer outdoor ambience with cicada chirping. [Listen to Ambience](https://retell-utils-public.s3.us-west-2.amazonaws.com/summer-outdoor.wav) - `mountain-outdoor`: Mountain outdoor ambience with birds singing. [Listen to Ambience](https://retell-utils-public.s3.us-west-2.amazonaws.com/mountain-outdoor.wav) - `static-noise`: Constant static noise. [Listen to Ambience](https://retell-utils-public.s3.us-west-2.amazonaws.com/static-noise.wav) - `call-center`: Call center work noise. [Listen to Ambience](https://retell-utils-public.s3.us-west-2.amazonaws.com/call-center.wav) Set to `null` to remove ambient sound from this agent. ' nullable: true ambient_sound_volume: type: number example: 1 description: If set, will control the volume of the ambient sound. Value ranging from [0,2]. Lower value means quieter ambient sound, while higher value means louder ambient sound. If unset, default value 1 will apply. language: oneOf: - $ref: '#/components/schemas/LanguageLegacy' - type: array items: $ref: '#/components/schemas/Language' description: Specifies what language(s) the agent will operate in. Accepts either a single scalar locale (e.g. `en-US`), the legacy scalar value `multi` for multilingual support, or an array of concrete locale codes for explicit multi-locale selection (e.g. `["en-US","es-ES"]`). The array form must contain concrete locale codes only — the `multi` value is valid only as the scalar legacy form and must not appear inside an array. Single-element arrays are normalized to the equivalent scalar on output. If unset, defaults to `en-US`. webhook_url: type: string example: https://webhook-url-here description: The webhook for agent to listen to call events. See what events it would get at [webhook doc](/features/webhook). If set, will binds webhook events for this agent to the specified url, and will ignore the account level webhook for this agent. Set to `null` to remove webhook url from this agent. nullable: true webhook_events: type: array items: type: string enum: - call_started - call_ended - call_analyzed - transcript_updated - transfer_started - transfer_bridged - transfer_cancelled - transfer_ended description: Which webhook events this agent should receive. If not set, defaults to call_started, call_ended, call_analyzed. nullable: true webhook_timeout_ms: type: integer example: 10000 description: The timeout for the webhook in milliseconds. If not set, default value of 10000 will apply. boosted_keywords: type: array items: type: string example: - retell - kroger description: Provide a customized list of keywords to bias the transcriber model, so that these words are more likely to get transcribed. Commonly used for names, brands, street, etc. nullable: true data_storage_setting: type: string enum: - everything - everything_except_pii - basic_attributes_only example: everything description: 'Granular setting to manage how Retell stores sensitive data (transcripts, recordings, logs, etc.). This replaces the deprecated `opt_out_sensitive_data_storage` field. - `everything`: Store all data including transcripts, recordings, and logs. - `everything_except_pii`: Store data without PII when PII is detected. - `basic_attributes_only`: Store only basic attributes; no transcripts/recordings/logs. If not set, default value of "everything" will apply. ' data_storage_retention_days: type: integer minimum: 1 maximum: 730 example: 30 nullable: true description: Number of days to retain call/chat data before automatic deletion. Must be between 1 and 730 days. If not set, data is retained forever (no automatic deletion). opt_in_signed_url: type: boolean example: true description: Whether this agent opts in for signed URLs for public logs and recordings. When enabled, the generated URLs will include security signatures that restrict access and automatically expire after 24 hours. signed_url_expiration_ms: type: integer example: 86400000 description: The expiration time for the signed url in milliseconds. Only applicable when opt_in_signed_url is true. If not set, default value of 86400000 (24 hours) will apply. nullable: true pronunciation_dictionary: type: array items: type: object required: - word - alphabet - phoneme properties: word: type: string example: actually description: The string of word / phrase to be annotated with pronunciation. alphabet: type: string enum: - ipa - cmu example: ipa description: The phonetic alphabet to be used for pronunciation. phoneme: type: string example: ˈæktʃuəli description: Pronunciation of the word in the format of a IPA / CMU pronunciation. description: A list of words / phrases and their pronunciation to be used to guide the audio synthesize for consistent pronunciation. Check the dashboard to see what provider supports this feature. Set to null to remove pronunciation dictionary from this agent. nullable: true end_call_after_silence_ms: type: integer example: 600000 description: If users stay silent for a period after agent speech, end the call. The minimum value allowed is 10,000 ms (10 s). By default, this is set to 600000 (10 min). max_call_duration_ms: type: integer example: 3600000 description: Maximum allowed length for the call, will force end the call if reached. The minimum value allowed is 60,000 ms (1 min), and maximum value allowed is 7,200,000 (2 hours). By default, this is set to 3,600,000 (1 hour). voicemail_message: type: string example: Hi, please give us a callback. description: The message to be played when the call enters a voicemail. Note that this feature is only available for phone calls. If you want to hangup after hitting voicemail, set this to empty string. voicemail_detection_timeout_ms: type: integer example: 30000 description: Configures when to stop running voicemail detection, as it becomes unlikely to hit voicemail after a couple minutes, and keep running it will only have negative impact. The minimum value allowed is 5,000 ms (5 s), and maximum value allowed is 180,000 (3 minutes). By default, this is set to 30,000 (30 s). voicemail_option: type: object properties: action: $ref: '#/components/schemas/VoicemailAction' detection_prompt: type: string maxLength: 2000 nullable: true description: Optionally describe what should be treated as voicemail. Leave as null to use the default definition. required: - action description: If this option is set, the call will try to detect voicemail in the first 3 minutes of the call. Actions defined (hangup, or leave a message) will be applied when the voicemail is detected. Set this to null to disable voicemail detection. example: action: type: static_text text: Please give us a callback tomorrow at 10am. nullable: true ivr_option: type: object properties: action: $ref: '#/components/schemas/IvrAction' detection_prompt: type: string maxLength: 2000 nullable: true description: Optionally describe what should be treated as an IVR. Leave as null to use the default definition. required: - action description: If this option is set, the call will try to detect IVR in the first 3 minutes of the call. Actions defined will be applied when the IVR is detected. Set this to null to disable IVR detection. example: action: type: hangup nullable: true call_screening_option: $ref: '#/components/schemas/CallScreeningOption' post_call_analysis_data: type: array items: $ref: '#/components/schemas/PostCallAnalysisData' description: Post call analysis data to extract from the call. This data will augment the pre-defined variables extracted in the call analysis. This will be available after the call ends. nullable: true post_call_analysis_model: $ref: '#/components/schemas/NullableLLMModel' example: gpt-4.1-mini description: The model to use for post call analysis. Default to gpt-4.1. analysis_successful_prompt: type: string maxLength: 2000 example: The agent finished the task and the call was complete without being cutoff. description: Prompt to determine whether the post call or chat analysis should mark the interaction as successful. Set to null to use the default prompt. nullable: true analysis_summary_prompt: type: string maxLength: 2000 example: Summarize the outcome of the conversation in two sentences. description: Prompt to guide how the post call or chat analysis summary should be generated. When unset, the default system prompt is used. Set to null to use the default prompt. nullable: true analysis_user_sentiment_prompt: type: string maxLength: 2000 example: Evaluate the user's sentiment based on their tone and satisfaction level. description: Prompt to guide how the post call or chat analysis should evaluate user sentiment. When unset, the default system prompt is used. Set to null to use the default prompt. nullable: true begin_message_delay_ms: type: integer example: 1000 description: If set, will delay the first message by the specified amount of milliseconds, so that it gives user more time to prepare to take the call. Valid range is [0, 5000]. If not set or set to 0, agent will speak immediately. Only applicable when agent speaks first. ring_duration_ms: type: integer minimum: 5000 maximum: 300000 example: 30000 description: If set, the phone ringing will last for the specified amount of milliseconds. This applies for both outbound call ringtime, and call transfer ringtime. Default to 30000 (30 s). Valid range is [5000, 300000]. stt_mode: type: string enum: - fast - accurate - custom example: fast description: If set, determines whether speech to text should focus on latency or accuracy. Default to fast mode. When set to custom, custom_stt_config must be provided. custom_stt_config: type: object description: Custom STT configuration. Only used when stt_mode is set to custom. properties: provider: allOf: - $ref: '#/components/schemas/AsrProvider' description: The STT provider to use. endpointing_ms: type: integer description: Endpointing timeout in milliseconds. Minimum is 100 for Azure, 10 for Deepgram, 500 for Soniox required: - provider - endpointing_ms nullable: true vocab_specialization: type: string enum: - general - medical example: general description: If set, determines the vocabulary set to use for transcription. This setting only applies for English agents, for non English agent, this setting is a no-op. Default to general. allow_user_dtmf: type: boolean example: true description: If set to true, DTMF input will be accepted and processed. If false, any DTMF input will be ignored. Default to true. allow_dtmf_interruption: type: boolean example: false description: If set to true, DTMF input will interrupt the agent even when interruption_sensitivity is 0. Can be overridden per conversation or subagent node. Default to false. user_dtmf_options: type: object properties: digit_limit: type: number description: The maximum number of digits allowed in the user's DTMF (Dual-Tone Multi-Frequency) input per turn. Once this limit is reached, the input is considered complete and a response will be generated immediately. nullable: true minimum: 1 maximum: 50 termination_key: type: string nullable: true description: A single key that signals the end of DTMF input. Acceptable values include any digit (0-9), the pound/hash symbol (#), or the asterisk (*). example: '#' timeout_ms: type: integer description: The time (in milliseconds) to wait for user DTMF input before timing out. The timer resets with each digit received. minimum: 1000 maximum: 15000 nullable: true denoising_mode: type: string enum: - no-denoise - noise-cancellation - noise-and-background-speech-cancellation example: noise-cancellation description: If set, determines what denoising mode to use. Use "no-denoise" to bypass all audio denoising. Default to noise-cancellation. pii_config: $ref: '#/components/schemas/PIIConfig' description: Configuration for PII scrubbing from transcripts and recordings. guardrail_config: $ref: '#/components/schemas/GuardrailConfig' description: Configuration for guardrail checks to detect and prevent prohibited topics in agent output and user input. handbook_config: $ref: '#/components/schemas/VoiceHandbookConfig' description: Toggle behavior presets on/off to influence agent response style and behaviors. timezone: type: string description: IANA timezone for the agent (e.g. America/New_York). Defaults to America/Los_Angeles if not set. example: America/New_York nullable: true is_public: type: boolean example: false description: Whether the agent is public. When set to true, the agent is available for public agent preview link. nullable: true ModelChoiceCascading: type: object required: - type - model properties: type: type: string enum: - cascading description: Type of model choice model: $ref: '#/components/schemas/LLMModel' description: The LLM model to use high_priority: type: boolean description: Whether to use high priority pool with more dedicated resource, default false Utterance: type: object required: - role - content - words properties: role: type: string enum: - agent - user - transfer_target description: Documents whether this utterance is spoken by agent or user. example: agent content: type: string description: Transcript of the utterances. example: hi how are you doing? words: type: array example: - word: hi start: 0.7 end: 1.3 description: Array of words in the utterance with the word timestamp. Useful for understanding what word was spoken at what time. Note that the word timestamp is not guaranteed to be accurate, it's more like an approximation. items: type: object properties: word: type: string description: Word transcript (with punctuation if applicable). start: type: number description: Start time of the word in the call in second. This is relative audio time, not wall time. end: type: number description: End time of the word in the call in second. This is relative audio time, not wall time. LLMModel: type: string enum: - gpt-4.1 - gpt-4.1-mini - gpt-4.1-nano - gpt-5 - gpt-5-mini - gpt-5-nano - gpt-5.1 - gpt-5.2 - gpt-5.4 - gpt-5.4-mini - gpt-5.4-nano - gpt-5.5 - claude-4.5-sonnet - claude-4.6-sonnet - claude-4.5-haiku - gemini-2.5-flash-lite - gemini-3.0-flash - gemini-3.1-flash-lite description: Available LLM models for agents. PIIConfig: type: object required: - mode - categories properties: mode: type: string enum: - post_call default: post_call description: The processing mode for PII scrubbing. Currently only post-call is supported. categories: type: array items: type: string enum: - person_name - address - email - phone_number - ssn - passport - driver_license - credit_card - bank_account - password - pin - medical_id - date_of_birth - customer_account_number uniqueItems: true default: [] description: List of PII categories to scrub from transcripts and recordings. ResponseEngineRetellLm: type: object required: - type - llm_id properties: type: type: string enum: - retell-llm description: type of the Response Engine. llm_id: type: string description: id of the Retell LLM Response Engine. version: type: number example: 0 description: Version of the Retell LLM Response Engine. nullable: true ResponseEngine: oneOf: - $ref: '#/components/schemas/ResponseEngineRetellLm' - $ref: '#/components/schemas/ResponseEngineCustomLm' - $ref: '#/components/schemas/ResponseEngineConversationFlow' ResponseEngineConversationFlow: type: object required: - type - conversation_flow_id properties: type: type: string enum: - conversation-flow description: type of the Response Engine. conversation_flow_id: type: string description: ID of the Conversation Flow Response Engine. version: type: number example: 0 description: Version of the Conversation Flow Response Engine. nullable: true VoicemailActionStaticText: type: object properties: type: type: string enum: - static_text example: static_text text: type: string example: Please give us a callback tomorrow at 10am. description: The text to be spoken when the call is detected to be in voicemail. required: - type - text GuardrailConfig: type: object properties: output_topics: type: array items: type: string enum: - harassment - self_harm - sexual_exploitation - violence - defense_and_national_security - illicit_and_harmful_activity - gambling - regulated_professional_advice - child_safety_and_exploitation uniqueItems: true description: Selected prohibited agent topic categories to check. When agent messages contain these topics, they will be replaced with a placeholder message. nullable: true input_topics: type: array items: type: string enum: - platform_integrity_jailbreaking uniqueItems: true description: Selected prohibited user topic categories to check. When user messages contain these topics, the agent will respond with a placeholder message instead of processing the request. nullable: true KBConfig: type: object properties: top_k: type: integer minimum: 1 maximum: 10 example: 3 description: Max number of knowledge base chunks to retrieve filter_score: type: number minimum: 0 maximum: 1 example: 0.6 description: Similarity threshold for filtering search results ResponseEngineCustomLm: type: object required: - type - llm_websocket_url properties: type: type: string enum: - custom-llm description: type of the Response Engine. llm_websocket_url: type: string description: LLM websocket url of the custom LLM. VoicemailAction: oneOf: - $ref: '#/components/schemas/VoicemailActionPrompt' - $ref: '#/components/schemas/VoicemailActionStaticText' - $ref: '#/components/schemas/VoicemailActionHangup' - $ref: '#/components/schemas/VoicemailActionBridgeTransfer' V2WebCallResponse: allOf: - type: object required: - call_type - access_token properties: call_type: type: string enum: - web_call example: web_call description: Type of the call. Used to distinguish between web call and phone call. access_token: type: string example: eyJhbGciOiJIUzI1NiJ9.eyJ2aWRlbyI6eyJyb29tSm9p description: Access token to enter the web call room. This needs to be passed to your frontend to join the call. - $ref: '#/components/schemas/V2CallBase' PostCallAnalysisData: oneOf: - $ref: '#/components/schemas/AnalysisData' - $ref: '#/components/schemas/CallPresetAnalysisData' description: Post-call analysis item (custom data or voice preset). Use for voice agent post_call_analysis_data; validates only call presets (call_summary, call_successful, user_sentiment). V2CallBase: type: object required: - call_id - agent_id - agent_version - call_status properties: call_id: type: string example: Jabr9TXYYJHfvl6Syypi88rdAHYHmcq6 description: Unique id of the call. Used to identify the call in the LLM websocket and used to authenticate in the audio websocket. agent_id: type: string example: oBeDLoLOeuAbiuaMFXRtDOLriTJ5tSxD description: Corresponding agent id of this call. agent_name: type: string example: My Agent description: Name of the agent. agent_version: type: integer example: 1 description: The version of the agent. call_status: type: string enum: - registered - not_connected - ongoing - ended - error example: registered description: 'Status of call. - `registered`: Call id issued, starting to make a call using this id. - `ongoing`: Call connected and ongoing. - `ended`: The underlying websocket has ended for the call. Either user or agent hung up, or call transferred. - `error`: Call encountered error. ' metadata: type: object description: An arbitrary object for storage purpose only. You can put anything here like your internal customer id associated with the call. Not used for processing. You can later get this field from the call object. retell_llm_dynamic_variables: type: object additionalProperties: {} example: customer_name: John Doe description: Add optional dynamic variables in key value pairs of string that injects into your Response Engine prompt and tool description. Only applicable for Response Engine. collected_dynamic_variables: type: object additionalProperties: {} example: last_node_name: Test node description: Dynamic variables collected from the call. Only available after the call ends. custom_sip_headers: type: object additionalProperties: type: string description: Custom SIP headers to be added to the call. example: X-Custom-Header: Custom Value data_storage_setting: type: string enum: - everything - everything_except_pii - basic_attributes_only example: everything description: Data storage setting for this call's agent. "everything" stores all data, "everything_except_pii" excludes PII when possible, "basic_attributes_only" stores only metadata. nullable: true opt_in_signed_url: type: boolean example: true description: Whether this agent opts in for signed URLs for public logs and recordings. When enabled, the generated URLs will include security signatures that restrict access and automatically expire after 24 hours. start_timestamp: type: integer example: 1703302407333 description: Begin timestamp (milliseconds since epoch) of the call. Available after call starts. end_timestamp: type: integer example: 1703302428855 description: End timestamp (milliseconds since epoch) of the call. Available after call ends. transfer_end_timestamp: type: integer example: 1703302628855 description: Transfer end timestamp (milliseconds since epoch) of the call. Available after transfer call ends. duration_ms: type: integer example: 10000 description: Duration of the call in milliseconds. Available after call ends. transcript: type: string example: 'Agent: hi how are you doing? User: Doing pretty well. How are you? Agent: That''s great to hear! I''m doing well too, thanks! What''s up? User: I don''t have anything in particular. Agent: Got it, just checking in! User: Alright. See you. Agent: have a nice day ' description: Transcription of the call. Available after call ends. transcript_object: type: array items: $ref: '#/components/schemas/Utterance' description: Transcript of the call in the format of a list of utterance, with timestamp. Available after call ends. transcript_with_tool_calls: type: array items: $ref: '#/components/schemas/UtteranceOrToolCall' description: Transcript of the call weaved with tool call invocation and results. It precisely captures when (at what utterance, which word) the tool was invoked and what was the result. Available after call ends. scrubbed_transcript_with_tool_calls: type: array items: $ref: '#/components/schemas/UtteranceOrToolCall' description: Transcript of the call weaved with tool call invocation and results, without PII. It precisely captures when (at what utterance, which word) the tool was invoked and what was the result. Available after call ends. recording_url: type: string example: https://retellai.s3.us-west-2.amazonaws.com/Jabr9TXYYJHfvl6Syypi88rdAHYHmcq6/recording.wav description: Recording of the call. Available after call ends. recording_multi_channel_url: type: string example: https://retellai.s3.us-west-2.amazonaws.com/Jabr9TXYYJHfvl6Syypi88rdAHYHmcq6/recording_multichannel.wav description: Recording of the call, with each party's audio stored in a separate channel. Available after the call ends. scrubbed_recording_url: type: string example: https://retellai.s3.us-west-2.amazonaws.com/Jabr9TXYYJHfvl6Syypi88rdAHYHmcq6/recording.wav description: Recording of the call without PII. Available after call ends. scrubbed_recording_multi_channel_url: type: string example: https://retellai.s3.us-west-2.amazonaws.com/Jabr9TXYYJHfvl6Syypi88rdAHYHmcq6/recording_multichannel.wav description: Recording of the call without PII, with each party's audio stored in a separate channel. Available after the call ends. public_log_url: type: string example: https://retellai.s3.us-west-2.amazonaws.com/Jabr9TXYYJHfvl6Syypi88rdAHYHmcq6/public_log.txt description: Public log of the call, containing details about all the requests and responses received in LLM WebSocket, latency tracking for each turntaking, helpful for debugging and tracing. Available after call ends. knowledge_base_retrieved_contents_url: type: string example: https://retellai.s3.us-west-2.amazonaws.com/Jabr9TXYYJHfvl6Syypi88rdAHYHmcq6/kb_retrieved_contents.txt description: URL to the knowledge base retrieved contents of the call. Available after call ends if the call utilizes knowledge base feature. It consists of the respond id and the retrieved contents related to that response. It's already rendered in call history tab of dashboard, and you can also manually download and check against the transcript to view the knowledge base retrieval results. latency: type: object description: Latency tracking of the call, available after call ends. Not all fields here will be available, as it depends on the type of call and feature used. properties: e2e: description: End to end latency (from user stops talking to agent start talking) tracking of the call. This latency does not account for the network trip time from Retell server to user frontend. The latency is tracked every time turn change between user and agent. $ref: '#/components/schemas/CallLatency' asr: description: Transcription latency (diff between the duration of the chunks streamed and the durations of the transcribed part) tracking of the call. $ref: '#/components/schemas/CallLatency' llm: description: LLM latency (from issue of LLM call to first speakable chunk received) tracking of the call. When using custom LLM. this latency includes LLM websocket roundtrip time between user server and Retell server. $ref: '#/components/schemas/CallLatency' llm_websocket_network_rtt: description: LLM websocket roundtrip latency (between user server and Retell server) tracking of the call. Only populated for calls using custom LLM. $ref: '#/components/schemas/CallLatency' tts: description: Text-to-speech latency (from the triggering of TTS to first byte received) tracking of the call. $ref: '#/components/schemas/CallLatency' knowledge_base: description: Knowledge base latency (from the triggering of knowledge base retrival to all relevant context received) tracking of the call. Only populated when using knowledge base feature for the agent of the call. $ref: '#/components/schemas/CallLatency' s2s: description: Speech-to-speech latency (from requesting responses of a S2S model to first byte received) tracking of the call. Only populated for calls that uses S2S model like Realtime API. $ref: '#/components/schemas/CallLatency' disconnection_reason: $ref: '#/components/schemas/DisconnectionReason' example: agent_hangup description: The reason for the disconnection of the call. Read detailed description about reasons listed here at [Disconnection Reason Doc](/reliability/debug-call-disconnect#understanding-disconnection-reasons). transfer_destination: type: string example: '+12137771234' description: The destination number or identifier where the call was transferred to. Only populated when the disconnection reason was `call_transfer`. Can be a phone number or a SIP URI. SIP URIs are prefixed with "sip:" and may include a ";transport=..." portion (if transport is known) where the transport type can be "tls", "tcp" or "udp". nullable: true call_analysis: description: Post call analysis that includes information such as sentiment, status, summary, and custom defined data to extract. Available after call ends. Subscribe to `call_analyzed` webhook event type to receive it once ready. $ref: '#/components/schemas/CallAnalysis' call_cost: description: Cost of the call, including all the products and their costs and discount. type: object required: - product_costs - total_duration_seconds - total_duration_unit_price - combined_cost properties: product_costs: type: array description: List of products with their unit prices and costs in cents items: $ref: '#/components/schemas/ProductCost' total_duration_seconds: type: number description: Total duration of the call in seconds example: 60 total_duration_unit_price: type: number description: Total unit duration price of all products in cents per second example: 1 combined_cost: type: number description: Combined cost of all individual costs in cents example: 70 llm_token_usage: type: object description: LLM token usage of the call, available after call ends. Not populated if using custom LLM, realtime API, or no LLM call is made. required: - values - average - num_requests properties: values: type: array items: type: number description: All the token count values in the call. average: type: number description: Average token count of the call. num_requests: type: number description: Number of requests made to the LLM. DTMFUtterance: type: object required: - role - digit properties: role: type: string enum: - dtmf description: Digit pressed by the user from their phone keypad. digit: type: string description: The digit pressed by the user. Will be a single digit string like "1", "2", "3", "*", "#" etc. example: '1' AgentVersionReference: oneOf: - type: integer minimum: 0 example: 1 - type: string minLength: 1 maxLength: 20 pattern: ^[a-z0-9_-]+$ example: prod description: Agent version reference. Supports a numeric version (for example 3) or a tag/environment name (for example "prod"). When a tag is provided, resolution uses that exact tag assignment (including its dynamic variables). If the tag exists but is currently unassigned, it resolves to latest. When a numeric version (or latest) is provided, resolution applies dynamic variables from the preferred tag for that resolved version (most recently assigned), if any. VoiceHandbookConfig: type: object description: Behavior presets for voice agents. All presets are available. properties: default_personality: type: boolean description: Professional call center rep baseline. natural_filler_words: type: boolean description: Sprinkle natural speech fillers like "um", "you know" for a more human, conversational tone. high_empathy: type: boolean description: Warm acknowledgment of caller concerns. echo_verification: type: boolean description: Repeat back and confirm important details (voice only). nato_phonetic_alphabet: type: boolean description: Spell using NATO phonetic alphabet style (voice only). speech_normalization: type: boolean description: Convert numbers/dates/currency to spoken forms (voice only). smart_matching: type: boolean description: Treat near-match similar words as same entity to reduce impact of transcription error (voice only). ai_disclosure: type: boolean description: When asked, acknowledge being a virtual assistant. scope_boundaries: type: boolean description: Stay within prompt/context scope, don't invent details. AsrProvider: type: string enum: - azure - deepgram - soniox description: ASR provider name. BooleanAnalysisData: type: object required: - type - name - description properties: type: type: string enum: - boolean description: Type of the variable to extract. example: boolean name: type: string description: Name of the variable. example: is_converted minLength: 1 description: type: string description: Description of the variable. example: Whether the customer converted. required: type: boolean description: Whether this data is required. If true and the data is not extracted, the call will be marked as unsuccessful. conditional_prompt: type: string description: Optional instruction to help decide whether this field needs to be populated in the analysis. If not set, the field is always included. If required is true, this is ignored. DisconnectionReason: type: string enum: - user_hangup - agent_hangup - call_transfer - voicemail_reached - ivr_reached - inactivity - max_duration_reached - concurrency_limit_reached - no_valid_payment - scam_detected - dial_busy - dial_failed - dial_no_answer - invalid_destination - telephony_provider_permission_denied - telephony_provider_unavailable - sip_routing_error - marked_as_spam - user_declined - error_llm_websocket_open - error_llm_websocket_lost_connection - error_llm_websocket_runtime - error_llm_websocket_corrupt_payload - error_no_audio_received - error_asr - error_retell - error_unknown - error_user_not_joined - registered_call_timeout - transfer_bridged - transfer_cancelled - manual_stopped VoicemailActionHangup: type: object properties: type: type: string enum: - hangup example: hangup required: - type StringAnalysisData: type: object required: - type - name - description properties: type: type: string enum: - string description: Type of the variable to extract. example: string name: type: string description: Name of the variable. example: customer_name minLength: 1 description: type: string description: Description of the variable. example: The name of the customer. examples: type: array items: type: string description: Examples of the variable value to teach model the style and syntax. example: - John Doe - Jane Smith required: type: boolean description: Whether this data is required. If true and the data is not extracted, the call will be marked as unsuccessful. conditional_prompt: type: string description: Optional instruction to help decide whether this field needs to be populated in the analysis. If not set, the field is always included. If required is true, this is ignored. RetellLlmOverride: type: object description: Override properties for Retell LLM configuration in agent override requests. properties: model: $ref: '#/components/schemas/NullableLLMModel' example: gpt-4.1 description: Select the underlying text LLM. If not set, would default to gpt-4.1. s2s_model: type: string enum: - gpt-realtime-1.5 - gpt-realtime - gpt-realtime-mini - null example: gpt-realtime-1.5 description: Select the underlying speech to speech model. Can only set this or model, not both. nullable: true model_temperature: type: number example: 0 description: If set, will control the randomness of the response. Value ranging from [0,1]. Lower value means more deterministic, while higher value means more random. If unset, default value 0 will apply. Note that for tool calling, a lower value is recommended. model_high_priority: type: boolean example: true description: If set to true, will use high priority pool with more dedicated resource to ensure lower and more consistent latency, default to false. This feature usually comes with a higher cost. nullable: true tool_call_strict_mode: type: boolean example: true description: Whether to use strict mode for tool calls. Only applicable when using certain supported models. nullable: true knowledge_base_ids: type: array items: type: string description: A list of knowledge base ids to use for this resource. nullable: true kb_config: type: object $ref: '#/components/schemas/KBConfig' description: Knowledge base configuration for RAG retrieval. nullable: true start_speaker: type: string enum: - user - agent description: The speaker who starts the conversation. Required. Must be either 'user' or 'agent'. begin_after_user_silence_ms: type: integer example: 2000 description: If set, the AI will begin the conversation after waiting for the user for the duration (in milliseconds) specified by this attribute. This only applies if the agent is configured to wait for the user to speak first. If not set, the agent will wait indefinitely for the user to speak. nullable: true begin_message: type: string example: Hey I am a virtual assistant calling from Retell Hospital. description: First utterance said by the agent in the call. If not set, LLM will dynamically generate a message. If set to "", agent will wait for user to speak first. nullable: true IvrAction: oneOf: - $ref: '#/components/schemas/IvrActionHangup' AnalysisData: oneOf: - $ref: '#/components/schemas/StringAnalysisData' - $ref: '#/components/schemas/EnumAnalysisData' - $ref: '#/components/schemas/BooleanAnalysisData' - $ref: '#/components/schemas/NumberAnalysisData' NumberAnalysisData: type: object required: - type - name - description properties: type: type: string enum: - number description: Type of the variable to extract. example: number name: type: string description: Name of the variable. example: order_count minLength: 1 description: type: string description: Description of the variable. example: How many the customer intend to order. required: type: boolean description: Whether this data is required. If true and the data is not extracted, the call will be marked as unsuccessful. conditional_prompt: type: string description: Optional instruction to help decide whether this field needs to be populated in the analysis. If not set, the field is always included. If required is true, this is ignored. NullableLLMModel: type: string enum: - gpt-4.1 - gpt-4.1-mini - gpt-4.1-nano - gpt-5 - gpt-5-mini - gpt-5-nano - gpt-5.1 - gpt-5.2 - gpt-5.4 - gpt-5.4-mini - gpt-5.4-nano - gpt-5.5 - claude-4.5-sonnet - claude-4.6-sonnet - claude-4.5-haiku - gemini-2.5-flash-lite - gemini-3.0-flash - gemini-3.1-flash-lite - null nullable: true description: Available LLM models for agents. ToolCallInvocationUtterance: type: object required: - role - tool_call_id - name - arguments properties: role: type: string enum: - tool_call_invocation description: This is a tool call invocation. tool_call_id: type: string description: Tool call id, globally unique. name: type: string description: Name of the function in this tool call. arguments: type: string description: Arguments for this tool call, it's a stringified JSON object. thought_signature: type: string description: Optional thought signature from Google Gemini thinking models. This is used internally to maintain reasoning chain in multi-turn function calling. IvrActionHangup: type: object properties: type: type: string enum: - hangup example: hangup required: - type AgentOverrideRequest: type: object description: Override configuration for agent, retell LLM, or conversation flow settings for a specific call. properties: agent: $ref: '#/components/schemas/AgentRequest' description: Override agent configuration settings. Any properties specified here will override the base agent configuration for this call. retell_llm: $ref: '#/components/schemas/RetellLlmOverride' description: Override Retell LLM configuration settings. Only applicable when using Retell LLM as the response engine. Supported attributes - model, s2s_model, model_temperature, model_high_priority, tool_call_strict_mode, knowledge_base_ids, kb_config, start_speaker, begin_after_user_silence_ms, begin_message. conversation_flow: $ref: '#/components/schemas/ConversationFlowOverride' description: Override conversation flow configuration settings. Only applicable when using conversation flow as the response engine. Supported attributes - model_choice, model_temperature, tool_call_strict_mode, knowledge_base_ids, kb_config, start_speaker, begin_after_user_silence_ms. EnumAnalysisData: type: object required: - type - name - description - choices properties: type: type: string enum: - enum description: Type of the variable to extract. example: enum name: type: string description: Name of the variable. example: product_rating minLength: 1 description: type: string description: Description of the variable. example: Rating of the product. choices: type: array items: type: string description: The possible values of the variable, must be non empty array. example: - good required: type: boolean description: Whether this data is required. If true and the data is not extracted, the call will be marked as unsuccessful. conditional_prompt: type: string description: Optional instruction to help decide whether this field needs to be populated in the analysis. If not set, the field is always included. If required is true, this is ignored. CallScreeningOption: type: object properties: agent_identity: type: string minLength: 1 maxLength: 100 example: Acme Health scheduling team description: Identity the agent should provide when a call screen asks who is calling. Dynamic variables are supported. call_purpose: type: string minLength: 1 maxLength: 300 example: confirming your appointment for tomorrow description: Purpose the agent should provide when a call screen asks why it is calling. Dynamic variables are supported. required: - agent_identity - call_purpose additionalProperties: false description: If this option is set, the agent prompt will include call screen handling instructions for identity and call purpose questions. Set this to null to disable call screen prompt instructions. nullable: true responses: UnprocessableContent: description: Unprocessable Content content: application/json: schema: type: object properties: status: type: string enum: - error message: type: string example: Cannot find requested asset under given api key. TooManyRequests: description: Too Many Requests content: application/json: schema: type: object properties: status: type: string enum: - error message: type: string example: Account rate limited, please throttle your requests. PaymentRequired: description: Payment Required content: application/json: schema: type: object properties: status: type: string enum: - error message: type: string example: Trial has ended, please add payment method. Unauthorized: description: Unauthorized content: application/json: schema: type: object properties: status: type: string enum: - error message: type: string example: API key is missing or invalid. BadRequest: description: Bad Request content: application/json: schema: type: object properties: status: type: string enum: - error message: type: string example: Invalid request format, please check API reference. InternalServerError: description: Internal Server Error content: application/json: schema: type: object properties: status: type: string enum: - error message: type: string example: An unexpected server error occurred. securitySchemes: api_key: type: http scheme: bearer bearerFormat: string description: Authentication header containing API key (find it in dashboard). The format is "Bearer YOUR_API_KEY"