New — Legible Connect

# Give Claude, Codex, Replit, and Lovable the right business context

Legible Connect lets AI tools pull fresh product, brand, FAQ, and document context from your Legible workspace — instead of relying on stale prompts or copied notes.

Works with Claude, Codex, Replit, Lovable, and other MCP-compatible tools.

Keep AI outputs aligned with your actual business

Stop re-pasting context into every session

One up-to-date knowledge layer across builders, agents, and assistants

What you get

# One connection. Better outputs everywhere.

Legible Connect turns your Legible knowledge layer into a live context service for AI tools. Your team can build pages, generate copy, write product flows, and answer customer questions with access to the same current business context.

Better quality

AI tools can retrieve current context instead of guessing from old prompts.

Faster workflows

Your team spends less time re-explaining the company and more time reviewing good output.

Fewer inconsistencies

The same source material powers builders, coding agents, and assistant experiences.

The real problem

# Most AI tools are smart. Most business context is still missing.

The problem is rarely model quality alone. The real issue is that generic AI tools don't know your pricing, positioning, product details, policies, support answers, or source documents — unless someone keeps feeding that context in manually. Legible Connect fixes that by letting the tool retrieve current context on demand.

How it works

# Connect in minutes

Built on top of the content Legible already keeps AI-ready. If you're already on Legible, you can connect in under a minute. If not, start with a free account and your content is indexed automatically.

1

Create an API key

Generate a Legible API key from your dashboard. Already on Legible? You're one click away.

2

Copy your endpoint

Grab your workspace MCP endpoint. It looks like: mcp.getlegible.com/{workspace_id}

3

Paste the MCP config

Add the endpoint and API key to your AI tool's MCP settings. One JSON block, 30 seconds.

4

Run the tester

Use the Legible Connect Tester in the dashboard to verify everything works end-to-end.

5

Start building

Your agent now retrieves your content automatically. Ask it anything about your business.

Architecture

# From content layer to context layer

The same content that powers your llms.txt, Markdown delivery, and Content Chat now powers your AI agents. One system, many outputs.

Your Website ──┐ FAQ Items ─────┼──▶ Legible Content Layer ──┬──▶ llms.txt / Markdown / AI Discovery Uploaded Docs ─┘ │ └──▶ Legible Connect (MCP) ──▶ Claude Code ──▶ OpenAI Codex ──▶ Custom Agents

MCP Tools

# Three tools your agent calls automatically

You don't need to learn new APIs. Your AI agent discovers these tools via MCP and calls them when it needs context.

search_context

Search your knowledge base for relevant content. Returns ranked chunks with source metadata.

get_document

Retrieve a full document by ID. Returns clean Markdown with metadata and frontmatter.

list_sources

List all content sources in your workspace with document counts and last-updated timestamps.

Use cases

# Every team benefits from AI that knows your business

Website generation

Let Replit, Lovable, or another builder pull your product context before it generates pages, sections, and messaging. No more re-explaining your business.

Coding agents

Give Claude Code or Codex access to your current business context while they build onboarding, support flows, docs, and product experiences.

Assistant quality

Use the same underlying knowledge layer across Content Chat, support integrations, and internal AI workflows. One source of truth, consistent answers.

Leadership & strategy

Draft investor updates, analyze positioning, and compare features — all grounded in real company data instead of generic AI output.

Why Legible Connect

# Designed for marketers and founders. Credible for developers.

Marketing and growth teams care that the output is faster, more accurate, and easier to manage. Developers care that the connection is clean, scoped, and based on structured retrieval instead of giant prompts. Legible Connect speaks to both.

Easy setup with a dedicated endpoint and API key
Structured tools instead of prompt stuffing
Fresh context as the business changes
One knowledge layer across customer-facing and internal AI use cases

Compatibility

Works with any MCP-compatible tool

Legible Connect uses the open Model Context Protocol. If your AI tool supports MCP, it works with Legible.

Claude Code
OpenAI Codex
Claude Desktop
Replit
Lovable
Cursor
Windsurf
Custom agents

FAQ

# Common questions

Do I need to build my own vector database integration?
No. Legible Connect sits on top of the knowledge layer Legible already generates and keeps current.
Does this replace the AI Export API?
No. Legible Connect is the simplest path for MCP-compatible tools. The AI Export API is still useful for custom sync pipelines and direct bulk export workflows.
What content can it use?
Website pages, FAQs, and uploaded documents that are already managed through Legible.
How do I know it's ready?
Run the Legible Connect Tester in the dashboard to verify the workspace, tools, sources, search, and document retrieval flow.

# Stop briefing your AI from scratch.

Connect your AI tools to the business context you already maintain in Legible and let them retrieve what they need, when they need it.