Bring your own keys.
Search.co never marks up model usage. You add your own provider keys; we encrypt them at rest (AES-256-GCM), never display them again, and route each agent task to the cheapest capable model across your keys — with automatic fallback if a provider fails.
Supported providers
OpenRouterOpenAIAnthropicGeminixAI (Grok)DeepSeekZ.aiGroqMistralTogetherCohereLocal endpointCustomAzure OpenAIBedrockVertexFireworksCerebrasSambaNovavLLM / Ollama / LM Studio
Encrypted at rest.
Keys are AES-256-GCM encrypted; only the last 4 characters are ever shown.
Routed by cost.
Cheap models for extraction, premium only for synthesis and risk — you set the policy.
Free by default.
No keys? Runs use a deterministic mock provider so you can try the whole flow at zero cost.