Documentation

Frequently Asked Questions

Common questions and answers about Fabric AI.

Find answers to the most common questions about Fabric AI.

General

What is Fabric AI?

Fabric AI is an enterprise platform that accelerates software development through intelligent AI agents, RAG-powered document generation, and durable workflow automation. It helps teams generate documents, automate repetitive tasks, and leverage their existing knowledge bases.

How is Fabric different from ChatGPT or Claude?

While ChatGPT and Claude are general-purpose AI assistants, Fabric AI is purpose-built for software development workflows:

FeatureGeneral AI AssistantsFabric AI
Document generationBasicStructured templates with RAG
External toolsLimited/NoneMCP integration (Jira, GitHub, Slack)
Workflow automationNoneTemporal-powered durable workflows
Team collaborationNoneMulti-tenant with role-based access
Context from your docsUpload each timePersistent workspaces with RAG

Is my data secure?

Yes. Fabric AI takes security seriously:

  • Data isolation: Multi-tenant architecture ensures your data is isolated
  • Encryption: All data encrypted at rest and in transit
  • API key security: Your AI provider keys are encrypted
  • No training: Your data is never used to train AI models
  • SOC 2 compliance: Enterprise-grade security practices

Which AI providers are supported?

Fabric AI supports all major AI providers through the Vercel AI SDK:

  • OpenAI (GPT-4, GPT-4 Turbo, GPT-3.5)
  • Anthropic (Claude 3.5 Sonnet, Claude 3 Opus, Claude 3 Haiku)
  • Google (Gemini Pro, Gemini Ultra)
  • Mistral
  • Cohere
  • And more

You can configure multiple providers and switch between them based on your needs.

Pricing & Plans

Is there a free tier?

Fabric AI offers a free tier with limited usage. You can upgrade to a paid plan for additional features and higher usage limits.

Do I need my own AI API key?

Yes. Fabric AI uses a BYOK (Bring Your Own Key) model. You provide your API keys for OpenAI, Anthropic, or other providers. This gives you:

  • Full control over costs
  • Choice of models
  • Direct relationship with providers
  • No markup on AI costs

Can I use Fabric with my team?

Yes! Create an organization to collaborate with your team:

  • Shared workspaces and documents
  • Organization-wide prompts and agents
  • Role-based access control
  • Centralized billing (optional)

Features

What is RAG and why does it matter?

RAG (Retrieval-Augmented Generation) allows AI agents to access your documents when generating responses. Instead of relying only on the AI's training data, agents can:

  • Reference your past PRDs and specs
  • Follow your company's templates
  • Use your terminology and conventions
  • Stay grounded in your actual data

Learn more about RAG →

What are MCP tools?

MCP (Model Context Protocol) is an open standard that lets AI agents interact with external tools. With MCP, Fabric agents can:

  • Create Jira tickets from PRDs
  • Post updates to Slack
  • Read and update GitHub issues
  • Access Notion databases
  • And many more integrations

Learn more about MCP →

What document types can Fabric generate?

The Document Generator can create:

  • Product Requirements Documents (PRDs)
  • Technical Specifications
  • Architecture Documents
  • API Documentation
  • User Stories
  • Release Notes
  • Meeting Summaries
  • Runbooks
  • And custom document types

Can I create custom agents?

Yes! You can create custom agents with:

  • Custom system prompts
  • Specific tool access
  • Custom approval workflows
  • Template variables
  • RAG context

Tutorial: Create Your First Agent →

Technical

How does the workflow engine work?

Fabric uses Temporal for durable workflow execution. This means:

  • Fault tolerance: Workflows resume after failures
  • Long-running: Support for workflows that take hours or days
  • Human-in-the-loop: Built-in approval workflows
  • Visibility: Real-time progress tracking

Learn more about Workflows →

What file types can I upload for RAG?

Fabric supports:

  • PDF (including scanned documents with OCR)
  • Microsoft Word (.docx)
  • Markdown (.md)
  • Plain text (.txt)
  • Various code files
  • HTML pages

Is there an API?

Yes, Fabric provides a comprehensive API for:

  • Managing agents and conversations
  • Uploading documents to workspaces
  • Triggering workflows programmatically
  • Accessing generated content

API Documentation →

Can I self-host Fabric?

Fabric offers multiple deployment options:

  • Cloud hosted: Fully managed at fabric.pro
  • Docker: Self-host with Docker Compose
  • .NET Aspire: Advanced local development

Installation Options →

Troubleshooting

My AI provider isn't working

  1. Verify your API key is correct
  2. Check that the provider is enabled in Settings
  3. Ensure you have credits/quota with the provider
  4. Test the connection in Settings → AI Gateway

Documents aren't uploading

  1. Check file format is supported
  2. Ensure file size is under 50MB
  3. Try refreshing the page
  4. Check your storage quota

Agent responses are slow

  1. Large context (many documents) takes longer to process
  2. Complex prompts require more computation
  3. Some models are slower than others
  4. Check your internet connection

MCP tools aren't available

  1. Verify the MCP server is configured in Settings
  2. Check that authentication is complete
  3. Test the server connection
  4. Ensure the agent has permission to use the tools

Workflow is stuck

  1. Check workflow status in the Workflows tab
  2. Look for pending approvals
  3. Check for failed activities
  4. Contact support if issue persists

Getting Help

Where can I get support?

  • Documentation: You're reading it!
  • Email: support@fabric.pro
  • Community: Join our Discord server
  • Enterprise: Dedicated support for enterprise customers

How do I report a bug?

  1. Go to Help → Report Issue
  2. Describe what you expected vs. what happened
  3. Include steps to reproduce
  4. Attach screenshots if helpful

Can I request a feature?

Yes! We love feature requests:

  1. Go to Help → Feature Request
  2. Describe your use case
  3. Explain how the feature would help
  4. Vote on existing requests

Still have questions?