Vapi MCP integration for AI agents.
Connect AI agents to Vapi through 38 structured actions, including create analytics query, create assistant, delete call, and test code tool execution. Revie…
- ACTION CONTRACTS
- 38
- AUTHENTICATION
- API key
- SOURCE STATUS
- Provider-backed
CATALOG SOURCE REVIEWED AUGUST 23, 2026 / ACTION NAMES AND SCHEMAS DERIVED FROM WORKING MACHINES PROVIDER SOURCE
CAPABILITY PROFILE
What the Vapi integration exposes
The Vapi Agent App exposes 38 provider-backed actions for ai and communication work. Its current contract lets an authorized agent create and execute one or more vapi analytics queries across call and subscription data, create a new vapi assistant with required transcriber, voice, and model settings plus optional messaging and duration controls, and delete a vapi call by its unique identifier. These operations are called through Working Machines as typed capabilities rather than through browser navigation or copied UI steps.
A connection uses API key. Before execution, the agent can inspect the selected action, its required fields, declared scopes, and expected output contract. Provider credentials remain inside the Working Machines runtime; the calling agent receives the capability and its structured result, not the underlying secret.
For reliable operation, start with the narrowest action that satisfies the task, resolve stable provider identifiers before changing state, and validate the returned object or status after execution. Availability still depends on the connected Vapi account, granted provider permissions, workspace policy, region, plan, and upstream API behavior.
VERIFIED ACTION SAMPLE
Real Vapi capabilities.
Showing 12 of 38 actions. Risk labels are conservative signals based on operation names, not substitutes for provider documentation or runtime policy.
create_analytics_queryREADCreate and execute one or more Vapi analytics queries across call and subscription data.
INPUTS: queries
create_assistantWRITECreate a new Vapi assistant with required transcriber, voice, and model settings plus optional messaging and duration controls.
INPUTS: name / model / voice / transcriber / firstMessage
delete_callHIGH IMPACTDelete a Vapi call by its unique identifier.
INPUTS: id
test_code_tool_executionREVIEWExecute TypeScript code inside Vapi's code tool sandbox and return the logs, result, and execution outcome.
INPUTS: code
create_evalWRITECreate a Vapi eval for a mock conversation and define the checkpoint messages used to evaluate model behavior.
INPUTS: name / type / messages / description
create_openai_chatWRITECreate an OpenAI-compatible Vapi chat response using an assistant or squad, with optional session and transport settings.
INPUTS: name / input / squad / stream / squadId
create_phone_numberWRITECreate a Vapi phone number using Vapi, Twilio, Vonage, Telnyx, or bring-your-own provider settings.
INPUTS: name / number / sipUri / provider / squadId
create_policyWRITECreate a Vapi monitoring policy with severity, threshold, and schedule or interval configuration.
INPUTS: name / interval / schedule / severity / threshold
create_provider_resourceWRITECreate a pronunciation dictionary provider resource in Vapi, defaulting to the 11labs pronunciation-dictionary route used by the upstream toolkit.
INPUTS: provider / resourceName / name / rules
create_scorecardWRITECreate a Vapi scorecard for observability and evaluation using structured output metrics and conditions.
INPUTS: name / metrics / description / assistantIds
create_sessionWRITECreate a Vapi session with either an assistant identifier or an inline assistant configuration.
INPUTS: name / assistant / assistantId
delete_chatHIGH IMPACTDelete a Vapi chat by its unique identifier.
INPUTS: id
INPUT CONTRACTS
Know what the action needs before it runs.
create_analytics_query
queriesREQUIRED- The analytics queries to execute.
create_assistant
nameOPTIONAL- Assistant identifier or display name.
modelREQUIRED- A raw JSON object returned by Vapi.
voiceREQUIRED- A raw JSON object returned by Vapi.
transcriberREQUIRED- A raw JSON object returned by Vapi.
firstMessageOPTIONAL- The first message the assistant should say.
delete_call
idREQUIRED- The unique identifier of the target resource.
test_code_tool_execution
codeREQUIRED- The TypeScript code that should be executed in Vapi's sandbox.
PROVIDER-SPECIFIC WORKFLOWS
Jobs this Agent App can support
Inspect Create Analytics Query
Create and execute one or more Vapi analytics queries across call and subscription data. Use this as a bounded discovery step, retain the returned identifier, and avoid expanding the read beyond the task's stated scope.
create_analytics_queryControl Create Assistant
Create a new Vapi assistant with required transcriber, voice, and model settings plus optional messaging and duration controls. Resolve the target first, present material changes for confirmation, and make retries idempotent where the provider supports it.
create_assistantVerify with Create Assistant
Create a new Vapi assistant with required transcriber, voice, and model settings plus optional messaging and duration controls. Compare the returned provider state with the intended outcome and preserve stable IDs or canonical links in the run record.
create_assistantSAFETY BOUNDARY
Operate Vapi with explicit limits
- Authorize Vapi with API key and grant only the provider access required by the selected actions.
- 19 actions are change-capable by name. Confirm the target identity and material parameters before allowing a write.
- Treat `delete_call`, `delete_chat`, and `delete_eval` as high-impact operations and require an explicit approval boundary.
- No provider scope string is declared on the sampled actions. Verify the connected account's actual permissions in Vapi rather than assuming unrestricted access.
- After a call, inspect the structured result and execution record before reporting that the Vapi task completed successfully.
CONNECTION MODEL
API key
Vapi API key used with the Authorization Bearer header.
Working Machines stores provider credentials behind the execution boundary. An agent can use an authorized connection identity, but catalog discovery alone does not reveal OAuth tokens, API keys, or provider secrets.
- READ SIGNALS
- 18
- WRITE SIGNALS
- 14
- HIGH IMPACT
- 5
- REVIEW SIGNALS
- 1
EVIDENCE AND AVAILABILITY
Provider reference
Action names, input fields, authentication types, and counts on this page are generated from the Working Machines provider catalog. Provider behavior, quotas, object semantics, account eligibility, and regional availability remain governed by Vapi.