# Coding Agents > Build ClutchCall apps with AI assistance — Claude Code, Cursor, Codex, and friends. The ClutchCall SDK is built to work well with AI coding agents. We ship three things designed specifically for them: 1. A documentation surface that LLMs can read directly (no HTML scraping). 2. A native MCP server they can call as a tool. 3. A CLI (`clutchcall`) that scaffolds projects with agent-ready `AGENTS.md` files baked in. If you're using Claude Code, Cursor, Codex CLI, or any other coding agent, point it at the resources below before asking it to write ClutchCall code. ## MCP server Available at: ``` https://docs.clutchcall.dev/api/mcp ``` Streamable HTTP transport (POST JSON-RPC). Three tools: | Tool | Args | Returns | |---------------|----------------------------|--------------------------------------------------------| | `list_pages` | (none) | Navigation-grouped slug index of every docs page. | | `search_docs` | `query`, `limit?` (≤25) | Ranked matches with title, description, and snippet. | | `get_page` | `slug` (e.g. `rpc/method-ids`) | Full Markdown body of the page. | ### Wiring into Claude Code, Cursor, etc. **Claude Code (.mcp.json):** ```json { "mcpServers": { "clutchcall-docs": { "url": "https://docs.clutchcall.dev/api/mcp" } } } ``` **Cursor (.cursor/mcp.json):** ```json { "mcpServers": { "clutchcall-docs": { "url": "https://docs.clutchcall.dev/api/mcp" } } } ``` **Codex CLI (codex.toml):** ```bash [mcp_servers.clutchcall-docs] url = "https://docs.clutchcall.dev/api/mcp" ``` Once configured, your agent will see `list_pages`, `search_docs`, and `get_page` as callable tools alongside its built-in toolbox. ### Smoke test ```bash curl -X POST https://docs.clutchcall.dev/api/mcp \ -H 'content-type: application/json' \ -d '{"jsonrpc":"2.0","id":1,"method":"tools/list"}' ``` ## Plain-text documentation For agents that don't speak MCP (or for use as a static prompt-context drop-in), every page is also published as plain Markdown: | URL | What it is | |--------------------------------------------------|-------------------------------------------------------------| | `https://docs.clutchcall.dev/llms.txt` | Short, navigation-grouped index. Triage starts here. | | `https://docs.clutchcall.dev/llms-full.txt` | Every page concatenated, page-marked. Whole-corpus context. | | `https://docs.clutchcall.dev/.md` | Per-page Markdown mirror (e.g. `/rpc/method-ids.md`). | | `https://docs.clutchcall.dev/search-index.json` | Searchable JSON corpus (powers the MCP `search_docs` tool). | The format follows the [llms.txt convention](https://llmstxt.org/) — drop the URL into any LLM's context window and it knows what to do with it. ## CLI: `clutch` The `clutch` CLI does what the MCP server does, plus runs telephony operations against the live control-plane API after `clutch auth login`. **Docs (no auth)**: ```bash clutch docs overview # navigation index clutch docs search "originate trunk" # full-text search clutch docs get-page rpc/method-ids # one page as Markdown ``` **Auth**: ```bash clutch auth login # browser OAuth → portal clutch auth whoami # current identity + active org ``` **Trunks + calls** (after auth): ```bash clutch trunks list # tabular view clutch trunks show # one trunk in detail clutch dial +15551234567 --trunk \ --app AI_BIDIRECTIONAL_STREAM --app-args my-agent-id clutch dial --list-apps # all 13 DialplanAction values clutch hangup ``` > **WARNING:** > `clutch dial` defaults `--app` to `AI_BIDIRECTIONAL_STREAM` (=6). If you > don't pass `--app-args ` the runtime has no agent to bridge > and the call lands in silence. For raw audio piping without an AI > agent, use `--app ANSWER` (=5). **Mid-call control verbs** (each calls `ExecuteDialplan` server-side): ```bash clutch transfer +15551234567 # RFC 3515 SIP REFER clutch mute # gateway-side TX silence clutch mute --wire # also send recvonly re-INVITE clutch unmute clutch hold clutch unhold clutch send-dtmf 5 --mode rfc2833 --duration-ms 200 ``` Override targets via env: `CLUTCH_BASE_URL` (portal), `CLUTCH_DOCS_BASE` (docs site). ## Scaffolding agent-ready projects `clutchcall init ` drops a working starter into your working directory: ```bash clutchcall init go my-voice-agent clutchcall init python my-py-agent clutchcall init typescript browser-demo ``` Each starter includes: - A working dial / event-handling sample (`main.clutchcall` / `main.py` / `src/index.ts`) that compiles against the ClutchCall SDK. - An **`AGENTS.md`** file tuned for that language. This is the file that agents read first when they enter the project — it captures the language-specific traps an agent will otherwise re-discover the hard way (e.g. "Emscripten corrupts byte arrays passed as `std::string`; use `typed_memory_view` instead"). - A `.clutchcall.json` project manifest with a pinned schema version for `clutchcall migrate`. ## Best practices for AI-assisted development These mirror what works for any rapidly-evolving SDK; we mention them because the ClutchCall wire format and method-ID table change between minor releases. 1. **Browse before searching.** Have your agent fetch `/llms.txt` first so it knows what exists. Search excerpts are noisier than full pages. 2. **Fetch full pages, not excerpts.** Use `get_page` (MCP) or `get-page` (CLI) when you need authoritative answers. The grouped index is for navigation; the page itself is for code. 3. **Never hand-edit serde structs.** The Python, TypeScript, Go, Rust, Java, and .NET bindings are auto-generated by `clutchcall/tools/apirpc_compiler/generate_bridges.py`. If your agent wants to add an RPC field, it should update the upstream schema and regenerate, not patch the binding by hand. 4. **Run `clutchcall migrate` after bumping the SDK.** Surfaces schema drift before runtime does. ## Agent Skills The [ClutchCall Agent Skills](https://github.com/clutchcall/agent-skills) repository ships Anthropic-format Skills — opinionated, end-to-end recipes that load into your coding agent's context when intent matches. Drop the repo into `.claude/skills/` (Claude Code) or `.cursor/skills/` (Cursor) and the agent gains task-specific knowledge for these workflows: **Generic recipes** - `clutchcall-pstn-dial` — outbound PSTN dialing (auth, originate, audio frames, hangup). - `clutchcall-ai-bridge` — wire a live call to OpenAI Realtime / Anthropic / Gemini Live with barge-in. - `clutchcall-call-events` — `CallEvent` stream consumption, reconnection, Q.850 cause handling. - `clutchcall-inbound-routing` — `SetInboundRouting` → `AnswerIncomingCall`, IVR menus, queueing. - `clutchcall-sip-trunk` — register a carrier (Twilio, Telnyx, Bandwidth, Plivo), caller-ID rules, codecs. - `clutchcall-bulk-campaign` — `OriginateBulk` with CSV input, pacing, abort + resume. - `clutchcall-recording` — per-call recording, stereo, retrieval, transcription. - `clutchcall-tool-calling` — LLM tool / function calling: the built-in implicit telephony toolset (transfer / hold / DTMF / disconnect / supervisor), custom server-side tools, and client-side tool calls via OpenAI Realtime. - `clutchcall-rag-integration` — pgvector / Pinecone / Weaviate retrieval as a tool call with latency budgets and barge-in safety. - `clutchcall-crm-enrichment` — pre-call enrichment (Salesforce / HubSpot lookup → system prompt overrides) vs. in-call tool lookup. **Vendor bridges** - `clutchcall-vendor-browser-webrtc` — capture mic in the browser, send PCMU over WebTransport. - `clutchcall-vendor-livekit` — bridge PSTN into a LiveKit room. - `clutchcall-vendor-daily` — bridge PSTN into a Daily room (Pipecat-friendly). - `clutchcall-vendor-chime` — Amazon Chime SDK Meetings or Voice Connector. - `clutchcall-vendor-twilio` — bridge from existing Twilio Media Streams (TwiML ``) into ClutchCall. - `clutchcall-vendor-vapi` — bridge a Vapi.ai voice agent over Vapi's WebSocket transport. Each skill is self-contained: prereqs checklist, working Python/TypeScript code, "verify it works" snippet, common gotchas, and pointers back to the docs MCP for deeper reference.