Provider Configuration

Connect your AI provider accounts to GateFlow. This guide covers supported providers and configuration options.

Supported Providers

Provider	Models	Status
OpenAI	GPT-5.2, GPT-5, o3, Whisper, Embeddings	Stable
Anthropic	Claude Opus 4.5, Claude Sonnet 4.5, Claude Haiku 4.5	Stable
Google	Gemini 3 Pro, Gemini 2.5 Pro, Gemini 2.5 Flash	Stable
Mistral	Large 3, Small 3, Voxtral, Embed	Stable
Cohere	Command R+, Command R, Embed, Rerank	Stable
ElevenLabs	Multilingual v2, Turbo v2.5, Flash v2.5	Stable

Adding a Provider

Via Dashboard

1. Go to Settings → Providers
2. Click Add Provider
3. Select the provider
4. Enter your credentials
5. Click Test Connection
6. Click Save

Via API

curl -X POST https://api.gateflow.ai/v1/management/providers \
  -H "Authorization: Bearer gw_prod_admin_key" \
  -H "Content-Type: application/json" \
  -d '{
    "provider": "openai",
    "credentials": {
      "api_key": "sk-..."
    },
    "settings": {
      "organization_id": "org-..."
    }
  }'

Provider-Specific Configuration

OpenAI

{
  "provider": "openai",
  "credentials": {
    "api_key": "sk-..."
  },
  "settings": {
    "organization_id": "org-...",
    "project_id": "proj_..."
  }
}

Available Models:

• gpt-5.2, gpt-5.2-chat-latest, gpt-5.2-codex
• gpt-5.1, gpt-5, gpt-5-mini, gpt-5-nano
• o3, o4-mini (reasoning models)
• text-embedding-3-large, text-embedding-3-small
• whisper-1, tts-1, tts-1-hd

Anthropic

{
  "provider": "anthropic",
  "credentials": {
    "api_key": "sk-ant-..."
  }
}

Available Models:

• claude-opus-4-5-20251107
• claude-sonnet-4-5-20250929
• claude-sonnet-4-20250514
• claude-haiku-4-5-20251015

Google (Gemini)

{
  "provider": "google",
  "credentials": {
    "api_key": "AIza..."
  }
}

Available Models:

• gemini-3-pro, gemini-3-flash
• gemini-2.5-pro, gemini-2.5-flash, gemini-2.5-flash-lite
• text-embedding-004

Mistral

{
  "provider": "mistral",
  "credentials": {
    "api_key": "..."
  }
}

Available Models:

• mistral-large-3, mistral-large-latest
• mistral-small-3, mistral-small-latest
• ministral-3b, ministral-8b, ministral-14b
• pixtral-large-latest
• devstral-2, devstral-small-2
• voxtral-mini-latest (speech-to-text)
• mistral-embed
• mistral-ocr-latest

Cohere

{
  "provider": "cohere",
  "credentials": {
    "api_key": "..."
  }
}

Available Models:

• command-r-plus, command-r-plus-08-2024
• command-r, command-r-03-2024
• embed-english-v3.0, embed-multilingual-v3.0
• rerank-english-v3.0, rerank-multilingual-v3.0

ElevenLabs

{
  "provider": "elevenlabs",
  "credentials": {
    "api_key": "..."
  }
}

Available Models:

• eleven_multilingual_v2 - Highest quality, 29 languages
• eleven_turbo_v2_5 - Low latency, optimized for real-time
• eleven_flash_v2_5 - Ultra-fast, cost-effective
• eleven_monolingual_v1 - English only, legacy

Voice IDs:

• rachel (21m00Tcm4TlvDq8ikWAM)
• josh (TxGEqnHWrfWFTfGW9XjX)
• bella (EXAVITQu4vr4xnSDxMaL)
• adam (pNInz6obpgDQGcFmaJgB)
• domi (AZnzlk1XvdvUeBnXmlld)

Testing Connections

Via Dashboard

Click Test Connection to verify credentials work.

Via API

curl -X POST https://api.gateflow.ai/v1/management/providers/openai/test \
  -H "Authorization: Bearer gw_prod_admin_key"

Response:

{
  "status": "ok",
  "latency_ms": 145,
  "models_available": ["gpt-5.2", "gpt-5", "..."]
}

Provider Priority

When multiple providers can serve a model, GateFlow uses priority order:

curl -X PATCH https://api.gateflow.ai/v1/management/providers/priority \
  -H "Authorization: Bearer gw_prod_admin_key" \
  -H "Content-Type: application/json" \
  -d '{
    "priority": ["openai", "anthropic", "google"]
  }'

Provider Health

GateFlow monitors provider health:

curl https://api.gateflow.ai/v1/management/providers/health \
  -H "Authorization: Bearer gw_prod_admin_key"

Response:

{
  "providers": [
    {
      "provider": "openai",
      "status": "healthy",
      "latency_p50_ms": 230,
      "latency_p99_ms": 890,
      "error_rate_1h": 0.001,
      "last_check": "2026-01-15T10:30:00Z"
    },
    {
      "provider": "anthropic",
      "status": "degraded",
      "latency_p50_ms": 450,
      "error_rate_1h": 0.05,
      "issues": ["Elevated latency detected"]
    }
  ]
}

Provider-Level Settings

Rate Limit Overrides

Set custom rate limits per provider:

{
  "provider": "openai",
  "settings": {
    "rate_limits": {
      "requests_per_minute": 500,
      "tokens_per_minute": 100000
    }
  }
}

Timeout Settings

{
  "provider": "openai",
  "settings": {
    "timeout_seconds": 60,
    "connect_timeout_seconds": 10
  }
}

Retry Configuration

{
  "provider": "openai",
  "settings": {
    "retry": {
      "max_attempts": 3,
      "initial_delay_ms": 1000,
      "max_delay_ms": 10000
    }
  }
}

Credential Security

Provider credentials are:

• Encrypted at rest with AES-256
• Never logged or exposed in responses
• Access controlled by organization roles
• Rotatable without downtime

Rotating Provider Credentials

1. Update credentials in the provider portal
2. Update in GateFlow:

curl -X PATCH https://api.gateflow.ai/v1/management/providers/openai \
  -H "Authorization: Bearer gw_prod_admin_key" \
  -H "Content-Type: application/json" \
  -d '{
    "credentials": {
      "api_key": "sk-new-key-here"
    }
  }'

3. Old credentials stop working immediately

Troubleshooting

"Provider Not Configured"

The model you requested requires a provider that isn't set up:

{
  "error": {
    "message": "Provider 'anthropic' not configured",
    "code": "provider_not_configured"
  }
}

Solution: Add the provider in Settings → Providers.

"Invalid Credentials"

Your provider API key is incorrect or expired:

{
  "error": {
    "message": "Provider authentication failed",
    "code": "provider_auth_error"
  }
}

Solution: Verify your API key in the provider's dashboard.

"Provider Rate Limit"

You've hit the provider's rate limit (not GateFlow's):

{
  "error": {
    "message": "Provider rate limit exceeded",
    "code": "provider_rate_limit",
    "retry_after_seconds": 60
  }
}

Solution: Wait and retry, or configure fallback models.

Next Steps

• Intelligent Routing - Route requests smartly
• Model Fallbacks - Handle failures gracefully