Skip to content

Kelki in your AI assistant

Just ask your AI assistant: “Find me a plumber in Créteil to fix a leak under the sink”. With the Kelki connector, ChatGPT, Claude or Le Chat prepares your request with you; once you confirm it by email, it is sent to local tradespeople in your area, and a professional calls you back.

Free · No obligation

Your request in 4 steps, without leaving the chat

  1. 01The tradeyour assistant suggests the trades that match your needs from across Kelki’s trades.
  2. 02Your need and timeframeyou choose from the same lists as the “Call me back” form on kelki.ai (services for that trade, or “Other need” in a few words), then the timeframe. You can add extra notes.
  3. 03Your townyour assistant checks the town and postcode.
  4. 04Your contact details and consentname, phone number and email address, a summary, then your consent to be called back.

One click in the email, and your request is sent

Right after, you receive an email from Kelki with your request summary and a “Confirm my request” button. One click, and your request goes to local trade pros in your area: only one pro calls you back to discuss it. The link remains valid for 48 hours; unconfirmed requests are automatically deleted, including contact details.

The service is completely free for you: tradespeople fund Kelki by paying 30 EUR excluding tax for the enquiries they take on. The estimate, appointment, and final pricing are then arranged directly with the professional.

Add Kelki to your assistant

Connector URL: https://kelki.ai/mcp — no account or password needed.

  • Claude: Customise › Connectors › “Add custom connector”, paste the URL, select “no authentication”.
  • ChatGPT: Plugins › “+” › custom MCP server, paste the URL, without authentication (depending on your ChatGPT plan).
  • Le Chat: (Mistral): Connectors › “Add a connector” › “Custom MCP connector”, paste the URL.
https://kelki.ai/mcp

Step‑by‑step guide, and setup in Claude Code, Cursor or VS Code: MCP Connector. On kelki.ai, WebMCP‑compatible browsers also offer the “demander_rappel” tool: it fills in the “Call me back” form on the page, which you submit yourself.

Kelki Agent: it finds the right pro for you

On every page of kelki.ai, the “Kelki Agent” bubble is here to help: “I have a leak”, “I’m locked out”, “I’m looking for a caterer”. It identifies the trade, offers services, timeframe and town as buttons, collects your details, then shows the summary: tick the consent box and send your request, just like the “Call me back” form.

Your own AI agent can also delegate the search via the A2A (Agent2Agent) protocol: it asks the same questions, then you receive the “Confirm my request” email. One click, and your request goes to local trade pros in your area.

Your data

The connector only receives what your assistant passes on for your request: trade, selected service (or your need in a few words), timeframe, town and postcode, your name, phone number, email address, any notes, and the name of the assistant used (e.g. “ChatGPT”). Your chat stays in your assistant: Kelki only receives these fields.

This information is used to send you the confirmation email, and then, once clicked, to route your request to local trade pros in your area. Any request not confirmed within 48 hours is deleted (leaving only trade, service, timeframe and town, with no contact details). Details, retention periods and rights: privacy policy.

For developers and platforms

  • Full connector documentation (the 4 tools, schemas, JSON‑RPC examples, setup): MCP Connector.
  • Remote MCP server: https://kelki.ai/mcp, Streamable HTTP transport, JSON responses, stateless, no authentication; protocol versions 2026‑07‑28, 2025‑11‑25, 2025‑06‑18 and 2025‑03‑26.
  • Tools: trouver_service, lister_prestations, verifier_lieu (read‑only) and demander_rappel (creates a request to confirm by email).
  • Resources: kelki://metiers, kelki://accord, kelki://metier/{metier}; prompt trouver_un_professionnel.
  • A2A Agent: https://kelki.ai/a2a, JSON‑RPC 2.0 (versions 1.0 and 0.3: SendMessage / message/send, SSE streams, GetTask, CancelTask), no authentication; task in input‑required on each question, with a data payload (step, choices, expected fields, summary, consent text); after consent, request to confirm by email (completed task). Agent card: /.well‑known/agent‑card.json.
  • Server card: /mcp/server‑card · ARD manifest: /.well‑known/ard.json · overview for assistants: /llms.txt.

Any questions?

Écrivez‑nous à contact@kelki.ai: nous répondons à chaque message. Kelki est édité par la société D.DESIGN & CONSEIL, à Paris.

Need a pro or have a question?
Get a call back.

Free · No obligation · Describe your need in a few clicks
Call me back
What you needEverywhere in FranceDescribe what you need · Free · No obligationCall me back