Welcome to Fabric¶
Fabric¶
Ship faster. Stay in control. Never fly blind.
Stop choosing between speed and quality. Fabric gives you AI coding superpowers while keeping you in the driver's seat.
Why Developers Love Fabric¶
You've tried AI coding tools. They're fast—until they hallucinate. They're helpful—until they miss context. They're powerful—until you lose control of your codebase.
Fabric is different. We built it for developers who want the velocity of AI without the anxiety of not knowing what's happening under the hood.
-
See AI Think, Not Just Output
Watch the AI reason through your problem in real-time. No more black boxes—understand why it suggests what it suggests, so you can trust (or correct) its decisions.
-
Choose the Right Model for the Job
Use fast models for quick tasks, powerful ones for complex reasoning. Switch mid-conversation. You're in control of the cost-quality tradeoff.
-
Work on Everything at Once
Juggle bug fixes, feature development, and code exploration in parallel. Each conversation keeps its context—no more losing your train of thought.
-
Your Code, Your Control
Every file change requires your approval. Full audit trail of AI actions. Your API keys, your data, your rules.
What Can You Build?¶
Ship Features in Hours, Not Days¶
Fabric doesn't just write code—it understands your codebase. Ask it to add a feature and it knows where things go, what patterns you use, and how to integrate properly.
Real Developer Workflows
- "Add JWT authentication to my Express API, following my existing middleware patterns"
- "Find all the places where we handle user permissions and add admin role support"
- "Refactor this class to use dependency injection—and update the tests"
Pick the Perfect Model for Every Task¶
Why pay premium prices for a simple rename? Why use a fast model for complex architecture decisions? Fabric lets you choose:
| When You Need... | Use | Why |
|---|---|---|
| Quick answers, simple fixes | Fast models | Instant response, minimal cost |
| Complex reasoning | Powerful models | Better architecture, fewer mistakes |
| Code review | Your choice | Match depth to importance |
Optimize Cost Without Sacrificing Quality
Switch models mid-conversation. Start with a fast model to explore, then switch to a powerful one when you need deep reasoning.
Let AI Handle the Boring Parts¶
Some tasks are tedious but important: updating tests after a refactor, applying a fix across 50 files, migrating to a new API. Fabric's agentic mode handles multi-step tasks while keeping you in control:
"Update all API endpoints to use the new error handling pattern,
and make sure the tests still pass."
You'll see exactly what it plans to do, approve each step, and watch it work. No surprises, no mystery commits, no "what did it just do?"
Your Codebase, Fully Understood¶
Fabric builds a mental model of your project. It knows:
- Where things live - Ask about auth code and it knows to look in
src/auth/ - How you do things - It learns your patterns, conventions, and style
- What connects to what - Understands relationships between files and modules
No More Explaining Your Codebase
You don't need to paste context into every prompt. Fabric remembers.
Beyond Code¶
Fabric isn't only for writing software. Because it combines a capable AI with a browser, a file system, and persistent context, it's just as useful for the thinking and research that surrounds the work.
Research Without Leaving Your Flow¶
The built-in browser lets Fabric read any page you're looking at and bring it directly into the conversation. Ask it to summarize a lengthy RFC, compare two libraries side by side, or pull the relevant section from a Stack Overflow thread — without copy-pasting anything.
Research Workflows
- "Read this Stripe docs page and show me the minimum setup I need for one-time payments"
- "Compare these two npm packages and tell me which is better maintained"
- "Summarize the key breaking changes in this migration guide"
Write Technical Docs That Actually Reflect the Code¶
Point Fabric at a module and ask it to write the README, the API reference, or the changelog entry. Because it reads the real code, it doesn't invent things — it describes what's actually there.
Writing Workflows
- "Write a README for this repo based on what the code actually does"
- "Generate a changelog entry from the commits since the last release"
- "Write a runbook for deploying this service, based on the scripts in /deploy"
Think Through Problems Before You Write a Line¶
Use Fabric as a thinking partner for architecture decisions, product tradeoffs, or incident post-mortems. Attach the relevant files and have a real conversation about the problem before committing to a direction.
Planning Workflows
- "I need to add real-time notifications to this app. Walk me through the tradeoffs of WebSockets vs SSE given what we already have"
- "We had an outage last night. Here's the log. Help me write a post-mortem"
- "I have to onboard a new engineer to this codebase next week. Help me write a walkthrough doc"
Get Started in 5 Minutes¶
-
1. Download
Available for macOS, Windows, and Linux. One-click install, no dependencies.
-
2. Connect Your AI
Add an API key from Anthropic, OpenAI, Google, or others. Free tiers available from Google, Mistral, and OpenRouter.
-
3. Open a Project
Point Fabric at your codebase and start asking questions. It learns your code as you go.
Free to Start
Don't want to pay for API access yet? Use free-tier providers like Google, Mistral, or OpenRouter.
What Will You Build Today?¶
Join the Community¶
-
Learn More
Dive deeper into Fabric's features with our guides.
-
Get Involved
Share feedback, request features, and help shape Fabric's future.
-
Power User Tips
Master keyboard shortcuts to fly through your workflow.