Management Chat API

The Management Chat API gives you Lexey over HTTP: the same agent as the Lexey pane in the portal, so you can configure your organization from code. It also has read-only endpoints for your knowledge, rules, configuration and customer conversations.

Authentication

All requests require a management API key (lxm_ prefix):

Authorization: Bearer lxm_...

Base URL

https://lexey.ai/api/v1/management

Quick start

Common tasks

Set business context:

{"message": "Our business is a SaaS company called Acme. We build project management tools for small teams."}

Add a knowledge article:

{"message": "Add a knowledge article about pricing: Our plans are Free (up to 3 projects), Pro ($12/month, unlimited projects), and Enterprise (custom pricing)."}

Update branding:

{"message": "Set our agent name to 'Acme Support', primary colour to '#2563EB', and welcome heading to 'Need help with Acme?'"}

Set business rules:

{"message": "Add these escalation rules: always escalate if a customer mentions billing disputes, refund requests over $100, or legal issues."}

Review analytics:

{"message": "How have we performed over the last 7 days?"}

Reading data directly

These endpoints return JSON and don't go through Lexey, so they're quick and use no credits. They use the same lxm_ key and base URL.

Paging. The two list endpoints return up to limit items (default 20, at most 100), with hasMore and nextCursor. Pass the cursor back as cursor_id plus cursor_updated_at (articles) or cursor_created_at (conversations):

Streaming

The send message endpoint returns an SSE stream, identical to the customer API. Each event includes a monotonically incrementing id: field and a data: line. For the full event format and code examples in TypeScript, Python, Go, Ruby, and C#, see the Streaming guide.

E2E testing

The management and customer APIs together enable end-to-end testing without browser automation:

1. Configure via management API — set up business context, knowledge, rules.
2. Test via customer API — send customer messages and verify agent responses.
3. Review via management API — check analytics, conversation history.

Limits and errors

• Messages to Lexey can be up to 20,000 characters, and a conversation up to 500 messages. After that the API returns 429 with code: "CONVERSATION_LIMIT_REACHED"; start a new conversation.
• A reply can take up to 300 seconds, because Lexey runs tools while it answers.
• Billing problems (no subscription, a past-due payment, credits used up) return JSON with status 402 before any stream starts, with an X-Billing-Reason header.
• Requests are rate limited; a 429 without the conversation-limit code means slow down and retry.

Building an LLM agent integration? See the Agent integration guide for a compact API reference.