Now in Private Preview

Ship better docs
in half the time

AI that reads your specs, respects your standards, and proposes changes you can review line-by-line.

Start Free See How It Works

No credit card required. Set up in 2 minutes.

See it in action

AI that understands your documents and implements improvements

app.contextloop.ai/projects/acme/documents/api-design

ACME Project

Project

Documents

getting-started.md

docs

api-design.md

architecture.md

specs

api-design.md

## API Design

Our REST API uses Express.js with TypeScript.

### Authentication

All endpoints require a valid JWT token.

Alice: Add rate limiting?1

```typescript

app.use(authMiddleware)

```

API Design

Our REST API uses Express.js with TypeScript.

Authentication

All endpoints require a valid JWT token.

app.use(authMiddleware)

AI Chat

Viewing: api-design.md

Inline Comments

Context-aware AI

AI Document Edits

The Problem with Current Tools

Generic AI

ChatGPT doesn't know your project context, codebase, or team decisions.

Siloed Docs

Changes just happen in Notion or Google Docs with no accountability or audit trail.

Lost Attribution

Who wrote this paragraph? Was it AI or a human? When was it changed?

Features

Everything you need for AI-powered documentation

A complete platform that combines writing, collaboration, and AI assistance—all with full transparency and control.

AI That Works in Context

Context-Aware AI

AI reads your project docs, specs, and connected tools. Answers grounded in your actual context.

8 Models, 5 Providers

Claude, GPT, Gemini, GLM, MiniMax. Switch models anytime. No vendor lock-in.

Extended Thinking

Toggle reasoning mode for complex problems. See AI's thought process with configurable effort levels.

Plan-Then-Act Agent

AI breaks down complex tasks into steps, explains its approach, and executes tools one at a time.

Agent Memory

AI remembers decisions, preferences, and gotchas from past conversations. Learning that persists.

Web Research

Connected to Exa for deep research. AI can search the web and bring insights into your docs.

Human Control Built In

Diff-First Edits

AI proposes changes as diffs. Accept or reject line-by-line. Feedback trains better suggestions.

Version Attribution

Every edit creates a version. Clear human vs AI attribution. Compare diffs, rollback anytime.

@Document References

Reference docs with @path/to/doc in chat. AI pulls context automatically. Comment threads anchor to sections.

Role-Based Access

Owner, Admin, Editor, Commenter, Viewer. Fine-grained permissions at project and document level.

Usage Analytics

Track AI token usage, costs per model, and team quotas. Full visibility into AI consumption.

Markdown-Native Docs

Rich Markdown Editor

Syntax highlighting for 100+ languages, Mermaid diagrams, KaTeX math. Plain text source of truth.

Semantic Search

Find anything with AI-powered search. Not just keywords—meaning.

PDF Export

Export any document as a polished PDF. Ready for sharing or printing.

Chat Attachments

Attach images, PDFs, and text files to chat. AI analyzes and responds with context.

Auto-Save

Never lose work. Changes saved automatically with full draft history.

Workflow

How ContextLoop Works

A simple workflow that keeps you in control while leveraging AI assistance.

Step 1

Create Project

Set up a workspace. Add documents, specs, and context your AI agents need.

Step 2

Add Context

Import specs, designs, and reference material. Connect research tools like Exa.

Step 3

Chat with AI

AI reads your context bundles to give grounded answers. No copy-pasting.

Step 4

Review Diffs

AI proposes a plan, then generates diffs. Accept or reject line-by-line.

Comparison

How We Compare

ContextLoop combines the best of documentation tools, AI assistants, and developer platforms.

Feature	ContextLoop	Notion AI	ChatGPT	Cursor
Context-aware AI AI understands your project docs, history, and team context
Document-centric AI chats Team members collaborate with AI in shared chat sessions tied to documents
AI diff proposals AI suggests changes as reviewable diffs you can accept or reject
Diff feedback loop AI learns from rejected suggestions to improve future proposals
Multi-model support Choose between different AI models (GPT, Claude, Gemini, etc.)
Markdown-native docs Documents stored as portable markdown, not proprietary formats
AI vs human attribution Version history distinguishes AI-generated from human-written content
CLI & API access Programmatic access via command-line tools and REST APIs
MCP server Exposes tools for external AI agents (Claude Code, Cursor) to connect
AI integrated comments AI participates in document comment threads alongside humans

For Developers

Built for developers, by developers

ContextLoop isn't just a web app. It's a platform with CLI, API, and integrations for your existing workflow.

CLI Integration

Upload, download, and manage docs from your terminal.

API Access

Full REST API with personal tokens. Build custom integrations.

Agent-Compatible

Works with Claude Code, Cline, and Cursor. Use your favorite tools.

MCP Server

Expose docs, context, and comments to Claude Code, Cursor, or Cline.

bash

# Sync a folder to your project (no global install)
npx contextloop-cli sync docs/ --project your-project-id

# Preview changes before syncing
npx contextloop-cli sync docs/ --project your-project-id --dry-run

# Or install globally
npm install -g contextloop-cli
contextloop sync docs/ --project your-project-id

Enterprise Ready

Security and compliance built-in

ContextLoop is built on enterprise-grade infrastructure with security at its core.

Row-Level Security

PostgreSQL RLS ensures data isolation. Users only see what they have access to.

Role-Based Access

Granular permissions from Owner to Viewer. Project and document-level controls.

Complete Audit Trail

Every action logged. Human vs AI attribution. Full version history.

Usage Metering

Track AI consumption per user, project, and model. Set quotas and limits.

Two-Factor Authentication

TOTP-based MFA with QR code enrollment. AAL2 security assurance level.

Data Portability

Markdown-native. Export anytime. No vendor lock-in on your content.

Pricing

Simple, Usage-Based Pricing

Pay only for what you use. No monthly subscriptions required.

AI Messages

Per-token usage, varies by model

Document Storage

Included up to quota

Team Members

Included up to tier limit

API Access

Included with account

Credit packages: $5, $10, $20, $50, $100

View Full Pricing

Context Bundles

Organize knowledge for AI

Context Bundles are curated collections of related documents. Toggle them on to give your AI exactly the context it needs—no more, no less.

API Guidelines

REST conventions, error formats, authentication patterns, and versioning rules. AI follows your API standards.

api-conventions.md

error-handling.md

auth-patterns.md

Architecture Docs

System diagrams, data models, and infrastructure specs. AI understands your architecture before suggesting changes.

system-overview.md

data-model.md

infrastructure.md

Style Guide

Writing tone, formatting rules, and terminology standards. AI matches your team's voice automatically.

writing-style.md

terminology.md

formatting-rules.md

AI can generate bundles automatically based on your project's content.

MCP Server

Your local agents, your team's context

ContextLoop exposes an MCP server that lets your favorite AI coding tools—Claude Code, Cursor, Cline—access your project's documents, context, and comments.

Read & Edit Documents

Agents can browse your project tree, read specs, and propose edits directly.

Access Project Context

Your curated context bundles are available to any connected agent.

Address Review Comments

Agents can read comments on specs, address feedback, reply, and resolve threads.

Team-Wide Visibility

All agent actions are tracked. Your team sees what changed and who (or what) changed it.

bash

# Add to your Claude Code MCP config
{
  "mcpServers": {
    "contextloop": {
      "command": "npx",
      "args": [
        "--yes",
        "contextloop-mcp@latest"
      ],
      "env": {
        "CONTEXTLOOP_PROJECT_ID": "your-project-id",
        "CONTEXTLOOP_API_KEY": "your-api-key"
      }
    }
  }
}

Works with Cursor, Cline, Windsurf, and any MCP-compatible agent.

Ready to give your docs an AI that actually gets it?

Join teams already using ContextLoop to write better documentation, faster.

Start Free Contact Sales

No credit card required

Set up in 2 minutes

Pay only for what you use

Feature

ContextLoop

Notion AI

ChatGPT

Cursor

Context-aware AI

AI understands your project docs, history, and team context

Document-centric AI chats

Team members collaborate with AI in shared chat sessions tied to documents

AI diff proposals

AI suggests changes as reviewable diffs you can accept or reject

Diff feedback loop

AI learns from rejected suggestions to improve future proposals

Multi-model support

Choose between different AI models (GPT, Claude, Gemini, etc.)

Markdown-native docs

Documents stored as portable markdown, not proprietary formats

AI vs human attribution

Version history distinguishes AI-generated from human-written content

CLI & API access

Programmatic access via command-line tools and REST APIs

MCP server

Exposes tools for external AI agents (Claude Code, Cursor) to connect

AI integrated comments

AI participates in document comment threads alongside humans

Built for developers, by developers

ContextLoop isn't just a web app. It's a platform with CLI, API, and integrations for your existing workflow.

CLI Integration

Upload, download, and manage docs from your terminal.

API Access

Full REST API with personal tokens. Build custom integrations.

Agent-Compatible

Works with Claude Code, Cline, and Cursor. Use your favorite tools.

MCP Server

Expose docs, context, and comments to Claude Code, Cursor, or Cline.

# Sync a folder to your project (no global install) npx contextloop-cli sync docs/ --project your-project-id # Preview changes before syncing npx contextloop-cli sync docs/ --project your-project-id --dry-run # Or install globally npm install -g contextloop-cli contextloop sync docs/ --project your-project-id

Your local agents, your team's context

ContextLoop exposes an MCP server that lets your favorite AI coding tools—Claude Code, Cursor, Cline—access your project's documents, context, and comments.

Read & Edit Documents

Agents can browse your project tree, read specs, and propose edits directly.

Access Project Context

Your curated context bundles are available to any connected agent.

Address Review Comments

Agents can read comments on specs, address feedback, reply, and resolve threads.

Team-Wide Visibility

All agent actions are tracked. Your team sees what changed and who (or what) changed it.

# Add to your Claude Code MCP config { "mcpServers": { "contextloop": { "command": "npx", "args": [ "--yes", "contextloop-mcp@latest" ], "env": { "CONTEXTLOOP_PROJECT_ID": "your-project-id", "CONTEXTLOOP_API_KEY": "your-api-key" } } } }

Ship better docsin half the time

See it in action

API Design

Authentication

The Problem with Current Tools

Generic AI

Siloed Docs

Lost Attribution

Everything you need for AI-powered documentation

AI That Works in Context

Human Control Built In

Markdown-Native Docs

How ContextLoop Works

Create Project

Add Context

Chat with AI

Review Diffs

How We Compare

Built for developers, by developers

CLI Integration

API Access

Agent-Compatible

MCP Server

Security and compliance built-in

Row-Level Security

Role-Based Access

Complete Audit Trail

Usage Metering

Two-Factor Authentication

Data Portability

Simple, Usage-Based Pricing

AI Messages

Document Storage

Team Members

API Access

Organize knowledge for AI

API Guidelines

Architecture Docs

Style Guide

Your local agents, your team's context

Read & Edit Documents

Access Project Context

Address Review Comments

Team-Wide Visibility

Ready to give your docs an AI that actually gets it?

Ship better docsin half the time

See it in action

API Design

Authentication

The Problem with Current Tools

Generic AI

Siloed Docs

Lost Attribution

Everything you need for AI-powered documentation

AI That Works in Context

Human Control Built In

Markdown-Native Docs

How ContextLoop Works

Create Project

Add Context

Chat with AI

Review Diffs

How We Compare

Built for developers, by developers

CLI Integration

API Access

Agent-Compatible

MCP Server

Security and compliance built-in

Row-Level Security

Role-Based Access

Complete Audit Trail

Usage Metering

Two-Factor Authentication

Data Portability

Simple, Usage-Based Pricing

AI Messages

Document Storage

Team Members

API Access

Ship better docs
in half the time

Ship better docs
in half the time