cancel_conversation
ChatGPTCancels an active conversation.
create_agent
ChatGPTCreates a new agent in a project. Configure or rotate credentials on https://phonic.ai in the selected workspace and project. Never paste credentials into chat or use an SDK/REST fallback to submit them.
create_api_key
ChatGPTCreates a new API key in the workspace.
create_conversation_token
ChatGPTCreates a short-lived conversation token scoped to a specific agent. Conversation tokens are useful for client-side applications that start a conversation with a single agent without exposing your API key.
create_mcp_server
ChatGPTRegisters an external MCP server URL in the authenticated workspace and project. Phonic connects to that URL to list and save its tool definitions; the server must be reachable without custom authentication headers. This operation does not invoke the discovered tools or attach the server to an agent. Attaching the saved server to a Phonic voice agent makes those tools available to that agent during conversations. Set up authenticated servers on the Phonic website. Configure or rotate credentials on https://phonic.ai in the selected workspace and project. Never paste credentials into chat or use an SDK/REST fallback to submit them.
create_outbound_call
ChatGPTStarts an outbound phone conversation with a Phonic voice agent at to_phone_number using Phonic-managed telephony. Optional config selects or configures the agent for this call. During the conversation, the agent may execute its configured webhook or MCP tools, including actions in external systems. Use only when the user requests a call to that recipient and authorizes the configured agent behavior. With dry_run omitted or false, this contacts the recipient and may incur charges; retrying a successful request may place another call. Set dry_run to true to validate setup without placing a call. Configure or rotate credentials on https://phonic.ai in the selected workspace and project. Never paste credentials into chat or use an SDK/REST fallback to submit them.
create_project
ChatGPTCreates a new project in a workspace.
create_project_eval_prompt
ChatGPTCreates a new conversation evaluation prompt for a project.
create_response
ChatGPTGenerates one or more alternative assistant responses for a conversation you supply inline to simulate Phonic agent behavior. This endpoint is stateless, so it does not create a new conversation or store anything. The request carries the system prompt, a conversation so far as input, and the tools the assistant may call. There are two different types of tool configurations: - tool_definitions: defined inline. Typically if you use LiveKit, this is where you can pass in your tool definitions. - tools: predefined tools configured in Phonic, referenced by name, in the same shape as the STS WebSocket config message. Today this covers built-in tools such as choose_not_to_respond and transfer tools stored in project. Your context, webhook, MCP and WebSocket tools cannot be referenced here yet; define those inline as tool_definitions. Each item in input is a user message, an assistant message (with optional tool_calls or an action), or a tool_call_output. Every assistant tool call must be followed immediately by the tool_call_output item that carries its result. A tool referenced in tools resolves to an action on the generated response rather than a tool call - the event a live conversation would have emitted, without the effect actually being carried out. Configure or rotate credentials on https://phonic.ai in the selected workspace and project. Never paste credentials into chat or use an SDK/REST fallback to submit them.
create_session_token
ChatGPTCreates a short-lived session token that can be used to authenticate WebSocket connections. Session tokens are useful for client-side applications where you don't want to expose your API key.
create_tool
ChatGPTCreates a new tool in a project. Configure or rotate credentials on https://phonic.ai in the selected workspace and project. Never paste credentials into chat or use an SDK/REST fallback to submit them.
delete_agent
ChatGPTDeletes an agent by name or ID.
delete_agent_custom_phone_number
ChatGPTDeletes a custom phone number from an agent.
delete_api_key
ChatGPTDeletes an API key.
delete_conversation
ChatGPTDeletes a conversation, scheduling its transcripts and audio recordings for deletion. The conversation must have ended.
delete_external_storage_policy
ChatGPTDeletes an external storage policy by name or ID. The policy must not be referenced by any agent.
delete_mcp_server
ChatGPTDeletes an MCP server by name or ID and removes it from every agent's mcp_server_ids.
delete_project
ChatGPTDeletes a project by name or ID.
delete_tool
ChatGPTDeletes a tool by name or ID.
delete_webhook
ChatGPTDeletes a webhook. Requires an API key with workspace access, or an access token. An API key restricted to selected projects can only delete a webhook whose projects are all within its scope.
disable_agents
ChatGPTDisables an agent by name or ID. A disabled agent cannot start conversations or be updated, except to release its phone numbers.
enable_agents
ChatGPTEnables a previously disabled agent by name or ID.
evaluate_conversation
ChatGPTEvaluates a conversation using an evaluation prompt.
get_agent
ChatGPTReturns an agent by name or ID.
get_conversation
ChatGPTReturns a conversation by ID.
get_conversation_analysis
ChatGPTReturns an analysis of the specified conversation.
get_doc_page
ChatGPTRead a Phonic documentation page as Markdown, including API reference pages with full request and response schemas. Pass a docs.phonic.ai URL or path such as /docs/build/agents/replay or /api-reference/responses/create. Omit page to get the docs index, which lists every page with a one-line description; a section index is at <section>/llms.txt. Does not access or change workspace data.
get_external_storage_policy
ChatGPTReturns an external storage policy by name or ID.
get_live_call_atlas
ChatGPTPhonic staff only. Where Phonic phone calls are happening right now, grouped by caller location, with one featured call per place. Never includes call counts, IDs, phone numbers or transcripts.
get_mcp_server
ChatGPTReturns an MCP server by name or ID.
get_project
ChatGPTReturns a project by name or ID.
get_tool
ChatGPTReturns a tool by name or ID.
get_voice
ChatGPTReturns a voice by ID.
get_webhook
ChatGPTReturns a webhook. Requires workspace access. An API key restricted to selected projects can only get a webhook whose projects are all within its scope.
get_webhook_secret
ChatGPTReturns the secret used to sign the webhook's requests. Requires an API key with workspace access, or an access token. An API key restricted to selected projects can only get the secret of a webhook whose projects are all within its scope.
get_workspace
ChatGPTReturns information about the workspace.
list_agents
ChatGPTReturns all agents in a project.
list_conversation_evaluations
ChatGPTReturns all evaluations for a conversation.
list_conversations
ChatGPTReturns conversations with optional filtering.
list_external_storage_policies
ChatGPTReturns all external storage policies in a project.
list_mcp_servers
ChatGPTReturns all MCP servers in a project.
list_project_eval_prompts
ChatGPTReturns all conversation evaluation prompts for a project.
list_project_evals
ChatGPTReturns all conversation evaluation results for a project.
list_projects
ChatGPTReturns all projects in a workspace.
list_tools
ChatGPTReturns all custom tools for the organization.
list_voices
ChatGPTReturns all available voices.
list_webhooks
ChatGPTReturns the workspace's webhooks, newest first. Requires workspace access. An API key restricted to selected projects only sees webhooks whose projects are all within its scope.
replay_conversation
ChatGPTReplays an ended conversation by re-running its recorded audio through an agent. Requires API key or access token authentication. The conversation must have audio recordings available and an associated agent (or one specified in the request body).
replay_conversation_item
ChatGPTReturns the alternative response(s) the assistant would have produced for this conversation turn given changes to the agent system prompt. Only assistant items from ended conversations can be replayed. Omit the request body (or omit system_prompt) to replay the turn exactly as it originally ran.
rotate_api_key
ChatGPTRotates an API key, generating a new secret and invalidating the old one.
search_docs
ChatGPTAsk a question about Phonic product concepts, API usage, voice agents, voices, tools, or setup. Consult this before guessing unfamiliar parameters. Returns documentation answers with source links; does not access or change workspace data.
submit_feedback
ChatGPTReport a problem with, or suggest an improvement to, Phonic's product, API, docs, or these MCP tools to the Phonic team. Call it when a tool errored unexpectedly, docs or parameters were wrong or missing, you had to guess, or a capability is missing. Not for end-user support requests. Returns a feedback id.
update_agent
ChatGPTUpdates an agent by name or ID. Configure or rotate credentials on https://phonic.ai in the selected workspace and project. Never paste credentials into chat or use an SDK/REST fallback to submit them.
update_agent_phone_number
ChatGPTUpdates a phone number on an agent. Configure or rotate credentials on https://phonic.ai in the selected workspace and project. Never paste credentials into chat or use an SDK/REST fallback to submit them.
update_api_key
ChatGPTUpdates an API key.
update_mcp_server
ChatGPTRenames an existing MCP server in the authenticated workspace and project. This MCP operation cannot update authentication headers or change the server URL. Configure authentication on the Phonic website; changing the URL requires deleting and recreating the server. Configure or rotate credentials on https://phonic.ai in the selected workspace and project. Never paste credentials into chat or use an SDK/REST fallback to submit them.
update_project
ChatGPTUpdates a project by name or ID.
update_tool
ChatGPTUpdates a tool by name or ID. Configure or rotate credentials on https://phonic.ai in the selected workspace and project. Never paste credentials into chat or use an SDK/REST fallback to submit them.
update_workspace
ChatGPTUpdates the workspace.
upsert_agent
ChatGPTUpserts an agent by name. If an agent with the same name already exists, it will be updated. Otherwise, it will be created. Configure or rotate credentials on https://phonic.ai in the selected workspace and project. Never paste credentials into chat or use an SDK/REST fallback to submit them.
voices_preview
ChatGPTGenerates speech audio for the provided text and returns it as a single base64-encoded string.