Blazing Agents

Blazing Agents runs the production infrastructure for your AI agents, so you can ship agent features without building a backend for them.

You write the agent logic: instructions, the model it uses, and the tools it can call. Blazing Agents runs everything around it. Conversations are stored, files survive between sessions, background tasks keep running after your request returns, and every turn's token usage is tracked. You call it from your backend with an SDK or plain HTTP. Read why teams use Blazing Agents for the longer case.

See it work

A typical integration is one chat route on your backend. Your frontend posts the user's message to it, and the route streams the agent's reply back. Set BLAZING_AGENTS_API_KEY from your dashboard and OPENROUTER_API_KEY from OpenRouter.

Run once: create the agent

Save a model provider and an agent that uses it. Run this script one time and keep the printed agent ID. Agent names are unique, so a second run fails; reuse the agent you already have.

import { BlazingAgents } from "@blazingagents/sdk";

const client = new BlazingAgents({
  apiKey: process.env.BLAZING_AGENTS_API_KEY!,
});

const provider = await client.providers.create({
  name: "OpenRouter",
  providerType: "openrouter",
  baseUrl: null,
  apiKey: process.env.OPENROUTER_API_KEY!,
});

const agent = await client.agents.create({
  name: "Support agent",
  instructions: "Answer clearly and briefly.",
  providerId: provider.id,
  model: "openai/gpt-6-luna",
});

console.log(`BLAZING_AGENTS_AGENT_ID=${agent.id}`);

Add the chat route

Set BLAZING_AGENTS_AGENT_ID to the printed ID. The route starts a new conversation when the request has no sessionId, and continues that conversation when it does.

import { BlazingAgents, type UIMessage } from "@blazingagents/sdk";

const client = new BlazingAgents({
  apiKey: process.env.BLAZING_AGENTS_API_KEY!,
});

export async function POST(request: Request): Promise<Response> {
  const { message, sessionId } = await request.json();
  const result = await client.chat({
    agentId: process.env.BLAZING_AGENTS_AGENT_ID!,
    message,
    sessionId,
  });
  return result.toResponse();
}

On the frontend, useChat with BlazingAgentsChatTransport renders the streamed reply and sends the session ID back on the next message; see Build a chatbot. Keep your API key on the backend, and before you continue a session, check that it belongs to the signed-in user. Connect Blazing Agents to your app covers that check.

What you can build

Start here

New to Blazing Agents:

1. Set up your API key and install an SDK.
2. Run your first agent.
3. Browse complete apps in the examples repository.

Find a task

Choose an interface