Kelki in your AI assistant
Ask your AI assistant: "Find me a plumber for a leak under the sink". With the Kelki connector, ChatGPT, Claude or Le Chat prepares your request with you; once you confirm by email, it goes out to local tradespeople across the 8 states and territories, and a professional calls you back.
Your request in 4 steps, without leaving the chat
- 01The tradeyour assistant suggests the trades that match your needs, chosen from the 1,000+ trades available on Kelki.
- 02Your job and timingchoose from the same options as the “Call me back” form on kelki.ai (trade services, or “Other need” in a few words), then how soon. You can add any extra details.
- 03Your suburbyour assistant checks the suburb and postcode.
- 04Your details and consentname, phone number and email, a quick summary, then your consent to receive a callback.
One click in the email, and your request is sent
Right after, you’ll receive an email from Kelki with a summary of your request and a “Confirm my request” button. One click, and your request goes out to trade pros in your local area: only one pro calls you back to discuss the job. The link is valid for 48 hours; unconfirmed requests are automatically deleted, including your contact details.
The service is completely free for you: tradespeople fund Kelki by paying 85 AUD excluding tax for the requests they accept. The quote, appointment and final price are then arranged directly with the professional.
Add Kelki to your assistant
Connector URL: https://kelki.ai/mcp — no account or password required.
- Claude: Customise › Connectors › “Add custom connector”, paste the URL, choose “no auth”.
- ChatGPT: Plugins › “+” › custom MCP server, paste the URL, no 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 setup, plus installation 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 page’s “Call me back” form for you to submit.
The Kelki Agent: finding the right pro for you
On every kelki.ai page, the “Kelki Agent” chat bubble is ready to help: “I’ve got a leak”, “I’m locked out”, “I need a caterer”. It works out the trade, gives you quick buttons for services, timing and your suburb, takes your details, and displays a summary: simply tick the consent box and send your request, just like the “Call me back” form.
Your own AI agent can also handle the search via the A2A (Agent2Agent) protocol: it asks the exact same questions, then you get the “Confirm my request” email. One click, and your request goes out to trade pros in your local area.
Your data
The connector only receives what your assistant sends for your request: the trade, the selected service (or your job description in a few words), the timing, the suburb and postcode, your name, phone number, email, any notes, and the name of the assistant used (e.g. “ChatGPT”). Your chat stays in your assistant: Kelki only receives these specific fields.
This information is used to send your confirmation email and, once you click, to pass your request on to trade pros in your local area. Requests not confirmed within 48 hours are deleted (retaining only the trade, service, timing and suburb, with no contact details). For details, retention periods and your rights, see our privacy policy.
For developers and platforms
- Full connector documentation (all 4 tools, schemas, JSON‑RPC examples, installation): 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 requiring email confirmation).
- 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 marked as input‑required at each step, with a data object (step, choices, expected fields, summary, consent text); once consented, email‑confirmation request created (completed task). Agent card: /.well‑known/agent‑card.json.
- Server card: /mcp/server‑card · ARD manifest: /.well‑known/ard.json · AI assistant summary: /llms.txt.
Got a question?
Drop us a line at contact@kelki.ai: we reply to every message. Kelki is published by D.DESIGN & CONSEIL, Paris.