ctalk
Enables coworkers to register profiles, ask questions, check inboxes, propose draft responses, and send approved replies through Claude Code.
README
ctalk
POC for pinging a coworker through Claude Code.
relay.py: hosted FastAPI + SQLite relay plus remote MCP endpoint.server.py: optional local stdio MCP bridge.
This is same-trust v0: no auth, no web UI, no push, no passwords.
Fast Remote MCP Setup
No local repo or Python install is required for the remote MCP path. Add the hosted MCP endpoint:
claude mcp add --transport http ctalk https://ctalk-relay.onrender.com/mcp -s user
Then register in Claude Code:
ctalk register me with user_id umar. My name is Umar. I own platform tooling.
Because this v0 has no auth, remote MCP prompts should include your user_id, such as umar or john.
Optional Local Stdio Install
If you want to run the MCP bridge locally instead of using hosted MCP:
cd /path/to/ctalk
python3 -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt
Deploy
The hosted app serves both the REST relay and the remote MCP endpoint at /mcp.
Render
Render's FastAPI docs use:
uvicorn relay:app --host 0.0.0.0 --port $PORT
Free-tier Render services do not support persistent disks. For the free tier, use:
Build Command: pip install -r requirements.txt
Start Command: uvicorn relay:app --host 0.0.0.0 --port $PORT
Environment: RELAY_DB_PATH=/tmp/relay.sqlite3
This works for a POC, but roster entries and messages may disappear after restarts or redeploys.
If you upgrade and add a persistent disk, this repo includes render.yaml with that start command plus a disk mounted at /var/data. Use:
Build Command: pip install -r requirements.txt
Start Command: uvicorn relay:app --host 0.0.0.0 --port $PORT
Environment: RELAY_DB_PATH=/var/data/relay.sqlite3
Disk Mount Path: /var/data
Without a persistent disk, do not use /var/data.
Railway
Railway can deploy this from GitHub or with the CLI. This repo includes railway.json with:
uvicorn relay:app --host 0.0.0.0 --port $PORT
Recommended Railway settings:
Start Command: uvicorn relay:app --host 0.0.0.0 --port $PORT
Healthcheck Path: /health
Environment: RELAY_DB_PATH=/data/relay.sqlite3
Volume Mount Path: /data
Public Networking: Generate Domain
After deployment, verify:
curl https://your-relay-url/health
Expected:
{"status":"ok"}
First Run
Each person registers once from Claude Code:
ctalk register me with user_id umar. My name is Umar. I own alpha-repo and platform tooling.
ctalk register me with user_id john. My name is John. I own alpha-repo and payments-service.
Send A Message
Ask:
ctalk ask john from umar: What is the retry logic in alpha-repo?
John checks:
ctalk check inbox for john.
John drafts:
ctalk propose a response as john to message 1: The retry logic uses three attempts with exponential backoff.
After John explicitly approves:
ctalk send the approved response as john to message 1: The retry logic uses three attempts with exponential backoff.
Umar checks replies:
ctalk check replies for umar.
Tools
register_self(user_id, name, ownership, notes = ""): creates your roster profile.ask_person(from_id, to_id, question): confirms the recipient exists, then posts a question.check_inbox(user_id): returns pending questions plus sender roster context.propose_response(user_id, message_id, draft_text): saves a draft only.send_response(user_id, message_id, final_text): delivers the answer after explicit human approval.check_replies(user_id, since = "1970-01-01T00:00:00+00:00"): polls answers.
HTTP API
GET /healthPOST /messageswith{ "from_id": "umar", "to_id": "john", "question": "..." }GET /inbox/{user_id}POST /messages/{id}/draftwith{ "draft_text": "..." }POST /messages/{id}/respondwith{ "final_text": "..." }GET /replies/{user_id}?since=<timestamp>POST /rosterwith{ "id": "john", "name": "John", "ownership": ["alpha-repo"], "notes": "" }GET /roster/{id}
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.