Integration

Voice.ai integration

Voice.ai provides text to speech with voice cloning and realtime AI phone agents. Generate speech audio from text, clone and manage voices, maintain pronunciation dictionaries, create, configure, deploy and pause voice agents, give them knowledge bases for retrieval, provision phone numbers, place outbound calls (approved accounts only), read call history with transcript and recording links, and connect an agent's Google account for its managed Calendar, Sheets and Gmail tools. Speech generation, cloning and calls use your Voice.ai plan. Not exposed: the chunked speech stream and the two WebSocket speech routes, and the realtime connection-details route (a session token for the client SDK). Authentication is a Voice.ai API key (dashboard, Developers) sent as Authorization: Bearer <key>.

Connect Voice.ai How integrations work

Ask it in chat

Say what should happen.
The workflow builds itself.

When this happens, do that in Voice.ai.Built from the 50 Voice.ai tools below, on the connection you add once.

What you can do with Voice.ai

Each of these is one operation on one endpoint, available both as a workflow step and as a tool an AI agent can call on its own. Nothing bundles several operations into one action, so what a run did stays readable afterwards.

Most endpoints run on the free plan. A few, mainly the ones that publish on your behalf, need a paid plan.

Knowledge base, 6 operations

create_knowledge_baseCreate knowledge basePOSTCreates a knowledge base from text documents for agents to search during calls; returns its kb_id (an existing one with identical content is returned instead). Attach it with assign_knowledge_base.
delete_knowledge_baseDelete knowledge baseDELETEDeletes a knowledge base.
get_knowledge_baseGet knowledge baseGETReturns one knowledge base with its documents.
list_knowledge_basesList knowledge basesGETLists your knowledge bases with name, description and document count, paginated as items and pagination.
unassign_knowledge_baseUnassign knowledge baseDELETEDetaches the knowledge base from an agent. The knowledge base itself is kept.
update_knowledge_baseUpdate knowledge basePUTUpdates a knowledge base's name or description, or its documents: documents sent REPLACE all existing ones. Leave documents out to keep them.

Agent, 4 operations

create_agentCreate agentPOSTCreates a voice agent from a name and a config (prompt, greeting, voice, model, webhooks); returns its agent_id. It takes no calls until deploy_agent.
get_agentGet agentGETReturns one agent: name, status, kb_id and its config (prompt, greeting, voice, model, webhooks, MCP servers). Credentials and signing secrets are not returned.
list_agentsList agentsGETLists your agents (deployed and paused by default; set show_statuses for others) with their config and status, paginated as items and pagination.
update_agentUpdate agentPUTUpdates an agent's name, config or knowledge base and returns the agent. Send only what changes: config fields you leave out are kept.

Pronunciation dictionary, 4 operations

delete_pronunciation_dictionaryDelete pronunciation dictionaryDELETEDeletes a pronunciation dictionary.
get_pronunciation_dictionaryGet pronunciation dictionaryGETReturns one pronunciation dictionary with its rules and versions.
list_pronunciation_dictionariesList pronunciation dictionariesGETLists your pronunciation dictionaries with id, name, language and current version.
rename_pronunciation_dictionaryRename pronunciation dictionaryPATCHRenames a pronunciation dictionary.

Voice, 4 operations

delete_voiceDelete voiceDELETEDeletes a voice you own.
get_voiceGet voiceGETReturns one voice: status, name and visibility. Public voices are readable by anyone signed in, private ones only by their owner.
list_voicesList voicesGETLists the voices you own (deleted ones excluded) with voice_id, name, status and visibility.
update_voiceUpdate voicePATCHRenames a voice you own or changes its visibility (PUBLIC makes it readable by any Voice.ai user).

Add from file, 1 operation

create_pronunciation_dictionary_from_fileCreate pronunciation dictionary from filePOSTCreates a pronunciation dictionary by importing a PLS (Pronunciation Lexicon) file; returns it with its rules.

Add from rule, 1 operation

create_pronunciation_dictionaryCreate pronunciation dictionaryPOSTCreates a pronunciation dictionary from a list of rules (word and replacement); returns it with its id for dictionary_id in generate_speech.

Add rule, 1 operation

add_pronunciation_rulesAdd pronunciation rulesPOSTAdds rules to a pronunciation dictionary (a new version); returns the dictionary.

Agent status, 1 operation

get_agent_connection_statusGet agent connection statusGETSays whether an agent can take a call now (call_allowed), with its status and the validation details.

Assign knowledge base, 1 operation

assign_knowledge_baseAssign knowledge basePOSTAttaches an existing knowledge base to an agent for retrieval during calls; returns the agent.

Available phone number, 1 operation

list_unassigned_phone_numbersList unassigned phone numbersGETLists the phone numbers you own that no agent uses yet.

Call history, 1 operation

list_call_historyList call historyGETLists past calls with agent, numbers, duration, status and credits, newest first per the vendor; filter by date range or agent_ids.

Clone voice, 1 operation

clone_voiceClone voicePOSTClones a voice from a reference audio file and returns its voice_id for generate_speech. Choose voice_visibility: PUBLIC voices are readable by any Voice.ai user.

Deploy, 1 operation

deploy_agentDeploy agentPOSTDeploys an agent so it can take calls (sets up its phone number when it has one); returns the deployed agent and a message.

Disable, 1 operation

delete_agentDelete agentPOSTDeletes an agent (its status becomes disabled). The agent must be paused first with pause_agent.

Disconnect, 1 operation

disconnect_googleDisconnect googleDELETEDisconnects the Google account linked to an agent.

Download, 1 operation

download_pronunciation_dictionaryDownload pronunciation dictionaryGETDownloads one version of a pronunciation dictionary as a PLS file.

End, 1 operation

end_callEnd callPOSTMarks a call as ended and returns its duration and credits used. The vendor meant it as a best-effort signal from the calling client.

Init agent, 1 operation

get_agent_templateGet agent templatePOSTReturns a pre-filled agent config for a template (customer_service, sales, support or custom) to pass to create_agent. It does not create an agent.

Model, 1 operation

list_modelsList modelsGETLists the LLM models (for an agent's llm_model) and TTS models (for model in generate_speech or tts_params) with their ids and names.

Outbound, 1 operation

create_outbound_callCreate outbound callPOSTPlaces a real phone call: the agent dials target_phone_number at once. Billed; restricted to approved accounts (403 otherwise). Returns call_id and status.

Pause, 1 operation

pause_agentPause agentPOSTPauses an agent and frees its phone number (idempotent); returns the paused agent. Needed before delete_agent.

Phone number, 1 operation

list_phone_numbersList phone numbersGETLists the phone numbers you own, with the agent each is assigned to and totals (total, available, assigned).

Recording, 1 operation

get_call_recording_urlGet call recording urlGETReturns a url to download the MP3 recording of one call (merged on the fly when not yet merged) and its status.

Release phone number, 1 operation

release_phone_numberRelease phone numberPOSTReleases a phone number you own, freeing its plan slot. A number attached to a deployed agent cannot be released: pause the agent first.

Remove rule, 1 operation

remove_pronunciation_rulesRemove pronunciation rulesPOSTRemoves rules from a pronunciation dictionary by rule id; returns the dictionary.

Search phone number, 1 operation

search_phone_numbersSearch phone numbersPOSTSearches phone numbers you could take with select_phone_number, by country and area code. Read-only.

Select phone number, 1 operation

select_phone_numberSelect phone numberPOSTProvisions a phone number from search_phone_numbers into your plan's allowance (one slot each). At the plan limit, release a number or upgrade first.

Set rule, 1 operation

replace_pronunciation_rulesReplace pronunciation rulesPOSTReplaces ALL rules of a pronunciation dictionary with the ones sent (a new version); returns the dictionary.

Speech, 1 operation

generate_speechGenerate speechPOSTGenerates speech from text and returns the audio file (mp3 by default). Uses voice_id when given, else a built-in voice. Billed per your Voice.ai plan.

Start, 1 operation

start_google_oauthStart google oauthPOSTStarts linking a Google account to an agent's managed tools: returns auth_url for a person to open and the requested scopes.

Stats summary, 1 operation

get_agent_status_summaryGet agent status summaryGETReturns the number of agents and a count per status.

Status, 1 operation

get_google_connection_statusGet google connection statusGETReturns whether an agent's Google account is connected, with granted, required and missing scopes and whether a reconnect is needed.

Test, 1 operation

test_events_webhookTest events webhookPOSTSends a test event to each enabled events webhook of the agent and returns the delivery result per endpoint.

Test inbound call, 1 operation

test_inbound_call_webhookTest inbound call webhookPOSTSends a sample inbound-call payload to the agent's inbound_call webhook and checks its answer (dynamic_variables, agent_overrides); returns the raw and validated response.

Test tool, 1 operation

test_tool_webhookTest tool webhookPOSTSends a sample function call to one tools webhook of the agent and returns the delivery result.

Transcript, 1 operation

get_call_transcript_urlGet call transcript urlGETReturns a url to download the transcript of one call.

Use Voice.ai from Claude Code, Cursor or Codex

Connect Voice.ai in LiveContext, create an API key in Settings, then add the LiveContext MCP server to your client. Its Voice.ai operations become tools the client can call, next to every other integration you connected.

Claude Code

claude mcp add --transport http livecontext https://livecontext.ai/mcp --header "X-API-Key: YOUR_API_KEY"

Codex (~/.codex/config.toml)

[mcp_servers.livecontext]
url = "https://livecontext.ai/mcp"

[mcp_servers.livecontext.http_headers]
"X-API-Key" = "YOUR_API_KEY"

Cursor, Claude Desktop and other JSON configs

{
  "mcpServers": {
    "livecontext": {
      "type": "http",
      "url": "https://livecontext.ai/mcp",
      "headers": {
        "X-API-Key": "YOUR_API_KEY"
      }
    }
  }
}

Set up the MCP server step by step

How to connect Voice.ai

Create your workspace
Sign up free. Nothing to install.
Connect Voice.ai once
Add your Voice.ai connection in Settings. Every tool from this integration reuses it, and secrets are encrypted at rest and visible only to you.
Describe the job in chat
The workflow builds itself in front of you and picks the endpoints it needs. Drag the steps yourself on the canvas whenever you would rather be explicit.