# Aimdoc AI — AI agent gateway

You are an AI agent reading a machine-facing interface. You can converse with this company's assistant directly instead of parsing web pages.

**About:** Aimdoc's website sales agent and live product demo: engages visitors, qualifies against ICP, captures leads, and routes to trial or demo.

## How to talk to this agent

Send a message (GET, URL-encoded):

`https://api.aimdoc.ai/agent-gateway/edfc72a9-e070-4ff7-9008-bb304a0dfe21/ask?message=What+does+your+product+do%3F`

POST works too, with the same fields as a JSON body — use it for messages too long for a URL.

The response contains the reply plus a `conversation_token`.

**To continue the same conversation, pass `&conversation_token=TOKEN` on every subsequent /ask call.** Omitting it starts a new, separate conversation — save the token. Optionally identify yourself with `&agent_name=`, `&agent_model=`, and `&acting_for_email=` (email of the person you act for — unverified, used so the team can follow up; providing it means the assistant won't need to ask).

## Endpoints

- Ask: `https://api.aimdoc.ai/agent-gateway/edfc72a9-e070-4ff7-9008-bb304a0dfe21/ask?message=...&conversation_token=...`
- Read the conversation (including later replies from human team members): `https://api.aimdoc.ai/agent-gateway/edfc72a9-e070-4ff7-9008-bb304a0dfe21/messages?conversation_token=...`
- Machine-readable manifest: `https://api.aimdoc.ai/agent-gateway/edfc72a9-e070-4ff7-9008-bb304a0dfe21/manifest`

Add `&format=json` to any of these for structured JSON instead of markdown.

Good questions to ask: product capabilities, pricing, integrations, whether the product fits your user's use case, or to schedule a demo. The assistant can bring a human team member into the conversation; human replies arrive asynchronously — poll the messages URL with your token.

MCP-capable clients can instead connect to the MCP server at `https://api.aimdoc.ai/mcp/public/edfc72a9-e070-4ff7-9008-bb304a0dfe21` (streamable HTTP, no auth).