FAQ

Find answers to common questions about Twig AI.

General

What is Twig AI?

Twig AI is an enterprise-grade RAG (Retrieval-Augmented Generation) platform that enables organizations to build, deploy, and manage intelligent AI agents powered by their own knowledge bases.

How does Twig AI work?

Twig AI combines your organization's data with advanced language models:

1. You connect data sources (documents, websites, databases)

2. Content is processed, chunked, and converted to vector embeddings

3. When users ask questions, relevant information is retrieved

4. The AI generates accurate responses using retrieved context

5. Responses include citations to source documents

What makes Twig AI different from ChatGPT?

• Your Data: Twig AI is trained on your specific knowledge base

• Citations: Every response includes sources for verification

• Privacy: Your data stays private and secure

• Control: Customize behavior, tone, and data access

• Enterprise Features: SSO, RBAC, audit logs, compliance

Is my data secure?

Yes. Twig AI employs enterprise-grade security:

• SOC 2 Type II compliant

• Data encrypted at rest and in transit

• Secure API key management

• SSO integration available

• Private data mode for sensitive information

• Regular security audits

Getting Started

How do I create my first agent?

1. Sign up at app.twig.so

2. Add a data source (upload files or connect integrations)

3. Create an agent and select your data sources

4. Test in the Playground

5. Deploy via API, widget, or extensions

See the Quick Start Guide for detailed steps.

What data sources can I connect?

Twig AI supports 14+ data source types:

• File uploads (PDF, Word, CSV, etc.)

• Websites and documentation

• Google Drive, OneDrive, SharePoint

• Confluence, Slack

• Zendesk

• And more

See Data Sources for the complete list.

How long does data processing take?

Processing time varies by source type and size:

• Small files (< 1MB): 1-2 minutes

• Websites (< 100 pages): 5-15 minutes

• Large integrations (1000+ documents): 30-60 minutes

You'll receive notifications when processing completes.

AI Agents

What's the difference between Redwood, Cedar, and Cypress?

These are RAG strategies with different speed-accuracy trade-offs:

• Redwood: Fastest (~1-2s), best for clear questions

• Cedar: Balanced (~2-3s), handles conversational queries

• Cypress: Highest accuracy (~3-4s), advanced reranking

See RAG Strategies for details.

Can I use multiple data sources per agent?

Yes! Agents can access any combination of data sources you configure. You can also restrict agents to specific sources or tags.

How do I improve answer accuracy?

1. Add more relevant data: Ensure knowledge base is comprehensive

2. Test and iterate: Use the Playground extensively

3. Review the Inbox: Mark responses as accurate/inaccurate

4. Choose the right strategy: Consider Cedar or Cypress for complex queries

5. Customize instructions: Provide clear behavior guidelines

6. Enable evals: Use the evaluation framework to track quality

Can agents use external information?

By default, agents can use both your private data and their general knowledge. You can enable "Private Data Only" mode to restrict agents to organization-specific information only.

Pricing & Billing

How is usage calculated?

Usage is measured by:

• Number of API requests

• Tokens processed (input + output)

• Data sources connected

• Storage used

Is there a free tier?

Yes! The free tier includes:

• 1,000 queries per month

• 3 data sources

• 1 agent

• Basic analytics

Upgrade to Pro or Enterprise for higher limits and advanced features.

What happens if I exceed my limits?

• API requests will return rate limit errors (429)

• You can upgrade your plan anytime

• Enterprise plans offer custom limits

Technical

What programming languages are supported?

Our REST API works with any language that can make HTTP requests. We provide official SDKs for:

• JavaScript/TypeScript ✅

• Python (coming soon)

• Ruby (coming soon)

Can I self-host Twig AI?

Twig AI is currently cloud-only. Enterprise customers can discuss dedicated deployments. Contact [email protected].

What's the API rate limit?

Rate limits vary by plan:

• Free: 20 requests/minute

• Pro: 100 requests/minute

• Enterprise: Custom limits

Does Twig AI support streaming responses?

Yes! Set stream: true in your API request to receive responses in real-time via Server-Sent Events (SSE).

Can I use custom language models?

Enterprise customers can bring their own models. Contact [email protected] to discuss custom deployments.

Privacy & Compliance

Where is my data stored?

Data is stored in secure cloud infrastructure with redundancy:

• Primary: AWS US-East

• Enterprise: Regional options available

Is Twig AI GDPR compliant?

Yes. Twig AI is fully GDPR compliant with:

• Data processing agreements

• Right to deletion

• Data portability

• Privacy by design

Can I delete my data?

Yes. You can delete:

• Individual data sources

• Specific documents

• Entire organizations

Deletions are permanent and cannot be undone.

Do you use my data to train your models?

No. Your data is never used to train language models or shared with other organizations. See our Privacy Policy for details.

Integrations

How do I embed the chatbot on my website?

1. Create an agent

2. Go to Widget settings

3. Customize appearance

4. Copy embed code

5. Add to your website's HTML

See Embeddable Widget for details.

Does Twig AI integrate with Zendesk?

Yes! Install the Zendesk app to get AI-powered suggestions directly in tickets. See Zendesk Integration.

Can I use Twig AI in Slack?

Yes! Install the Slack app to:

• Ask questions in channels

• Get AI responses via direct message

• Share answers with team

See Slack App for setup.

Support

How do I get help?

• Documentation: This help center

• Email: [email protected]

• Chat: In-app chat widget

• Community: community.twig.so

What's your support response time?

• Free: Best effort (24-48 hours)

• Pro: 12-hour response time

• Enterprise: 4-hour response time, dedicated support

Do you offer implementation help?

Yes! Enterprise customers receive:

• Dedicated customer success manager

• Implementation guidance

• Training sessions

• Custom integrations support

Contact [email protected] to learn more.

Troubleshooting

My data source isn't syncing

1. Check process logs for errors

2. Verify authentication credentials

3. Ensure URLs are accessible

4. Check file format compatibility

5. Contact support if issues persist

The AI can't answer my questions

1. Verify relevant information is in your knowledge base

2. Check data source is connected to your agent

3. Try rephrasing your question

4. Review cited sources in the response

5. Consider adding more data sources

Responses are too slow

1. Check your internet connection

2. Consider using Redwood strategy (faster)

3. Reduce topK if set very high

4. Verify system status at status.twig.so

API requests are failing

1. Verify API key is correct and active

2. Check rate limits haven't been exceeded

3. Ensure proper request format (JSON)

4. Review error message and code

5. Check API documentation

---

Still have questions? Contact us at [email protected]