Build with the models you love.
Let Gutman handle the connections, routing and control that bring them together.
BRING YOUR PROVIDERS. KEEP YOUR OPTIONS OPEN.
Go from scattered integrations to a clear view of your AI. Connect providers, decide how requests flow, and understand what happens next.
Give your application a stable model alias. Choose targets by priority, cost or latency, and configure fallback providers for eligible failures before a response starts streaming.
Issue virtual keys for projects and applications. Control model access, set request limits and expiry, and revoke access without sharing your upstream credentials.
Follow routing attempts, latency and token usage. Review recorded costs and supported provider charges, so you can make better decisions about your stack.
Visibility that turns usage into understanding.Power chat and tool workflows through a unified API. Swap configured model targets while keeping the same route in your application.
Chat · Streaming · ToolsConnect media workflows through fal.ai and BytePlus, and voice through ElevenLabs. Manage provider access alongside your text models.
Image · Video · AudioTurn your text and Markdown documents into searchable knowledge collections. Bring relevant context into AI workflows with controlled collection access.
Embeddings · Search · RAGChoose a unified route for flexible model selection, or a provider-native key to keep your provider’s SDK and request format.
One endpoint for supported chat and embedding workflows, with routing and fallback controls.
Use original model IDs and supported native APIs through a scoped provider key. Native requests keep provider semantics and do not add route failover.
// Your stack. One entry point.
import OpenAI from "openai";
const ai = new OpenAI({
apiKey: process.env.GUTMAN_API_KEY,
baseURL: "https://api.gutman.ai/v1",
});
const response = await ai.chat.completions.create({
// The route alias you configured in Gutman
model: "support-assistant",
messages: [{
role: "user",
content: "Let's build something great."
}],
});// Your provider's SDK. Your virtual key.
import OpenAI from "openai";
const ai = new OpenAI({
apiKey: process.env.GUTMAN_NATIVE_KEY,
baseURL: "https://api.gutman.ai/native/openai",
});
const response = await ai.responses.create({
// An original model ID from your provider
model: process.env.OPENAI_MODEL_ID,
input: "Let's build something great.",
});Yes. Connect your own provider accounts and API credentials. Gutman manages access and routing; model availability, subscriptions and provider charges remain with your provider.
For supported integrations, yes. Use the unified endpoint with an OpenAI-compatible client, or a provider-native key with the proxy URL shown in the console. Compatibility depends on the provider and operation.
Configure a route alias and its model targets, then select a priority, cost or latency policy. Eligible failures can move to another configured target before streaming begins. Native provider keys do not add automatic fallback.
Gutman records usage and available cost information, including supported token estimates and provider-reported charges. Coverage varies by provider and operation. Estimates are not a replacement for your provider’s invoice.
Put your AI stack on a foundation you control.
Open the Gutman console ↗