agentvoice
agentvoice is an MCP server for real phone calls: your coding agent can call you when it gets stuck, or place disclosed calls to businesses.
This repo is the source of the @spekoai/mcp-calls
npm package. Calls run on Speko; every call opens with an
AI disclosure, and safety rails (business-line checks, rate caps, local DNC, after-hours
confirmation, signed dial tokens) are enforced server-side.
Quickstart
# sign in with your browser, then configure every detected supported client
npx @spekoai/mcp-calls@latest init
Then tell your agent:
> "work on <task> and call me when you are done or stuck"
mcp/README.md is the canonical user doc: setup for every client, the tool reference, the CLI, safety rails, and hermetic test mode.
Architecture
Default (single process):
MCP host -> stdio MCP + embedded server core -> @spekoai/sdk -> api.speko.dev
key, owner/DNC state, lookup credentials, and rails live here
Optional remote mode:
MCP host -> stdio MCP -> authenticated HTTP -> Express server -> @spekoai/sdk -> api.speko.dev
keys, state, lookup, and rails live here
mcp/— the stdio MCP and CLI. In the default single-process install it embeds the server core and readsSPEKO_API_KEYfrom the client's MCP environment.server/— the reusable calling core plus optional Express wrapper. It runs lookup, owner resolution, safety rails, and@spekoai/sdkdialing.
The source is split so the same trusted core can run embedded or behind Express. Google lookup
stays out of api.speko.dev; it belongs to this app and runs wherever the core runs.
Layout
mcp-dev-calls/
├── mcp/ # stdio MCP + CLI; embeds the core by default (mcp/README.md = user doc)
├── server/ # reusable trusted core + optional Express wrapper + safety-rail tests
├── characterization/ # golden-baseline suite driven by test:integration
├── docs/ # guides — agent-platforms.md: running on cloud agent platforms
├── scripts/ # check-version-lockstep.mjs (CI gate), mcp-e2e.mjs (integration),
│ # place-call.mjs (one-shot call runner), inspect-call.mjs (diagnostics)
├── .env.example # both tiers
└── package.json # npm workspaces root
Develop
npm run typecheck # tsc --noEmit, both tiers
npm test # vitest, both tiers
npm run build # tsc → dist/, both tiers
npm run test:integration # characterization + MCP e2e + selftest
npm run dev:server # backing API with watch
npm run dev:mcp # MCP over stdio with watch
MIT © SpekoAI
