Frequently Asked Questions

General

What is GateFlow?

GateFlow is an AI Gateway that sits between your application and AI providers (OpenAI, Anthropic, Google, etc.). It provides a unified API, automatic fallbacks, semantic caching, and compliance features.

How is GateFlow different from using providers directly?

Feature	Direct API	GateFlow
Multi-provider	Integrate each separately	Single API for all
Fallbacks	Build yourself	Automatic
Caching	Build yourself	Semantic caching included
Cost tracking	Multiple dashboards	Unified analytics
Model deprecation	Manual migration	Automatic fallbacks

Is GateFlow open source?

GateFlow is a managed service. We don't currently offer self-hosted options, but Enterprise customers can discuss on-premise deployment.

Pricing & Billing

Do you add markup to provider costs?

No. We pass through provider costs at the same rate. You pay exactly what you would pay the provider directly.

How do I control costs?

• Set daily/monthly budgets per API key or organization
• Use semantic caching to reduce duplicate calls
• Configure cost alerts
• Use cheaper models for simpler tasks

Are cached responses free?

Yes! When a response is served from cache, you pay nothing for that request.

Technical

Which providers do you support?

• OpenAI (GPT-4o, GPT-4, GPT-3.5, Whisper, Embeddings)
• Anthropic (Claude 3.5, Claude 3)
• Google (Gemini 1.5 Pro, Gemini 1.5 Flash)
• Mistral (Large, Medium, Small, Codestral)
• Cohere (Command R+, Embeddings, Rerank)

Is GateFlow compatible with the OpenAI SDK?

Yes! Just change the base URL:

from openai import OpenAI

client = OpenAI(
    base_url="https://api.gateflow.ai/v1",
    api_key="gw_prod_..."
)

What's the latency overhead?

Typically 10-50ms for non-cached requests. Cached requests return in <50ms total.

Do you support streaming?

Yes, full streaming support for all providers that offer it.

What happens if a provider goes down?

GateFlow automatically routes to your configured fallback models. You can configure fallback chains per model.

Security & Compliance

Is my data stored?

• Request/response logs are stored for analytics (configurable retention)
• Semantic cache stores response content (configurable TTL)
• You can disable logging and caching per request

Do you support HIPAA?

Yes, Enterprise plans include BAA signing and HIPAA-compliant data handling.

Where is data stored?

By default, data is stored in US regions. Enterprise customers can configure data residency for EU or other regions.

Are provider API keys secure?

Provider API keys are encrypted at rest using AES-256 and are never logged or exposed in responses.

API Keys

What do the key prefixes mean?

• gw_dev_ - Development keys (for testing)
• gw_prod_ - Production keys (for live applications)
• gf-agent- - MCP agent keys

Can I restrict what a key can do?

Yes, you can restrict keys to:

• Specific models
• Rate limits (requests/minute)
• Cost limits (daily/monthly)
• IP addresses

How do I rotate keys?

1. Create a new key
2. Update your application
3. Revoke the old key

We also support automatic rotation with webhooks.

Troubleshooting

"Invalid API Key" error

• Ensure your key starts with gw_prod_ or gw_dev_
• Check for extra whitespace
• Verify the key hasn't been revoked

"Model not found" error

• Check the model name matches exactly (e.g., gpt-4o not gpt4o)
• Verify the provider is configured in your dashboard
• Check your API key has permission to use that model

High latency

• Enable semantic caching
• Check if you're hitting rate limits (causes queuing)
• Try a different provider/model
• Contact support if consistently slow

Unexpected costs

• Review analytics to see which models/keys are driving costs
• Enable cost alerts
• Set budget limits
• Increase cache usage

Getting Help

How do I contact support?

• Email: support@gateflow.ai
• Dashboard: Click "Help" in the sidebar
• Enterprise: Dedicated Slack channel

Is there a status page?

Yes: status.gateflow.ai

Where can I report bugs or request features?

• GitHub Issues: github.com/gateflow/gateflow/issues
• Email: feedback@gateflow.ai