Cloud Providers
Cloud providers each show up as a card in Settings → Models. Click a card, paste your API key (and a region or model list if the provider needs one), and click Save. The provider's models appear in the model selector immediately.
You only need an API key for the providers you actually use — Synapse picks the provider for each call from the model prefix.
Quick reference
| Provider | UI card label | Required fields |
|---|---|---|
| Anthropic | Anthropic Claude | API Key |
| OpenAI | OpenAI | API Key |
| Google Gemini | Google Gemini | API Key |
| xAI Grok | xAI Grok | API Key |
| DeepSeek | DeepSeek | API Key |
| AWS Bedrock | AWS Bedrock | API Key, AWS Region, (optional) Inference Profile |
| Cloud OpenAI-compatible (OpenRouter, Together, …) | OpenAI Compatible | API Key, Base URL, Model Names |
For local providers (Ollama, vLLM, LM Studio), see Local Providers.
Anthropic (Claude)
Model prefix: claude-
- Open Settings → Models and click the Anthropic Claude card.
- Paste your API key into API Key (format:
sk-ant-...). The card links to the Anthropic Console if you need to generate one. - Click Save. The card's status dot turns green and Claude models appear in the model selector.
| Model | Best for |
|---|---|
claude-opus-4-7 | Most capable — complex reasoning, long context |
claude-sonnet-4-6 | Balanced — quality + speed |
claude-3-5-sonnet-20241022 | Strong general purpose |
claude-haiku-4-5-20251001 | Fast and cheap — routing, classification |
Advanced: direct settings.json edit
{ "anthropic_key": "sk-ant-..." }
OpenAI (GPT)
Model prefixes: gpt-, o1-, o3-
In Settings → Models click the OpenAI card, paste your key (format sk-...) into API Key, click Save.
| Model | Best for |
|---|---|
gpt-4o | Flagship — vision + text |
gpt-4o-mini | Fast and cheap |
o1 | Deep reasoning |
o3-mini | Fast reasoning |
Advanced: direct settings.json edit
{ "openai_key": "sk-..." }
Google Gemini
Model prefixes: gemini-, gemma-
In Settings → Models click the Google Gemini card and paste your key (format AIza...) into API Key. The card links to Google AI Studio where you can get a free key.
| Model | Best for |
|---|---|
gemini-2.0-flash | Fast, large context |
gemini-1.5-pro | Multimodal, 1M token context |
gemma-3-27b-it | Open-weight via Gemini API |
Advanced: direct settings.json edit
{ "gemini_key": "AIza..." }
xAI (Grok)
Model prefix: grok-
In Settings → Models click the xAI Grok card and paste your key (format xai-...) into API Key.
| Model | Description |
|---|---|
grok-3 | Latest Grok model |
grok-2-vision | Multimodal |
Advanced: direct settings.json edit
{ "grok_key": "xai-..." }
DeepSeek
Model prefix: deepseek-
In Settings → Models click the DeepSeek card and paste your key (format sk-...) into API Key.
| Model | Description |
|---|---|
deepseek-chat | General purpose (DeepSeek-V3) |
deepseek-reasoner | Deep reasoning (DeepSeek-R1) |
Advanced: direct settings.json edit
{ "deepseek_key": "sk-..." }
AWS Bedrock
Model prefix: bedrock.
Bedrock exposes Anthropic, Meta, Mistral, and Amazon foundation models through one AWS endpoint. In Settings → Models click the AWS Bedrock card and fill:
| Field | Example | Notes |
|---|---|---|
| Bedrock API Key | absk-... or bedrock-api-key... | Generate in AWS Console — the card links to the docs |
| AWS Region | us-east-1 | Region where your Bedrock access is enabled |
| Inference Profile (Optional) | us.anthropic.claude-3-5-sonnet-20241022-v2:0 | Required for cross-region inference. Synapse populates a dropdown from your account; pick one or leave blank for direct invocation |
Models are referenced with the full Bedrock model ID prefixed by bedrock.:
bedrock.anthropic.claude-3-5-sonnet-20241022-v2:0
bedrock.meta.llama3-8b-instruct-v1:0
bedrock.mistral.mistral-large-2402-v1:0
Embedding model for Bedrock mode
When you operate Synapse entirely in Bedrock mode (no Gemini/OpenAI keys), set an embedding model for long-term memory in the same card:
- Embedding Model:
amazon.titan-embed-text-v2:0
Advanced: direct settings.json edit
{
"bedrock_api_key": "absk-...",
"aws_region": "us-east-1",
"bedrock_inference_profile": "us.anthropic.claude-3-5-sonnet-20241022-v2:0",
"embedding_model": "amazon.titan-embed-text-v2:0"
}
OpenAI-Compatible (Cloud)
For cloud providers that expose an OpenAI-compatible API — OpenRouter, Together AI, Fireworks AI, Anyscale, etc.
Model prefix: oaic.<model_name>
In Settings → Models click the OpenAI Compatible card and fill:
| Field | Example |
|---|---|
| API Key | sk-or-... (provider-specific) |
| Base URL | https://openrouter.ai/api (or your provider's v1 base) |
| Model Names (comma-separated) | meta-llama/llama-3-70b-instruct, google/gemma-2-27b-it |
| Embedding Model Names (comma-separated, optional) | hf:nomic-ai/nomic-embed-text-v1.5 |
Then use the model in agent settings as oaic.meta-llama/llama-3-70b-instruct (prefix + the name you listed).
Advanced: direct settings.json edit
{
"openai_compatible_key": "sk-or-...",
"openai_compatible_base_url": "https://openrouter.ai/api/v1",
"openai_compatible_models": "mistralai/mistral-7b,meta-llama/llama-3-70b-instruct",
"openai_compatible_embed_models": "text-embedding-3-small"
}
Prompt caching
All cloud providers above benefit from the prompt cache, which is on by default. You'll see savings reflected in Settings → Usage without any per-provider configuration.