One API endpoint that researches, writes documents, executes code, processes files — and calls your functions. Bring your own model. We handle the rest.
const result = await stewrd.agent.runWithTools( { message: "Check order #4821 and email a shipping update", tools: [getOrder, sendEmail], }, async (call) => await handlers[call.name](call.arguments))// "Done — shipping update sent to [email protected]"Orchestration, sandboxing, file storage, tool execution, streaming — managed. Your LLM key handles the thinking. Stewrd handles the doing.
Pick what your project needs. Each request uses only the capabilities you specify.
3 credits
Web search, page scraping, and multi-source synthesis. Returns reports, not links.
5 credits
Generate PDFs, spreadsheets, and presentations from your data.
3 credits
Execute Python and Node.js in a sandboxed environment. Process files, run analysis.
const output = await run("python")1 credit
Parse, transform, and analyze structured data — CSV, JSON, Excel.
One unified API
Pick what you need per project. A request using multiple capabilities costs the sum of their weights.
From API key to shipping in minutes.
// Dashboard → New Project// Select: Research, Documents// Copy your API keysk-stw_a1b2c3d4e5f6...200 free credits every month. Buy more when you need them.
Credits are weighted by capability: Chat/Data = 1, Files = 2, Research/Code = 3, Documents = 5. Custom tool output submissions are free. Purchased credits never expire.
Real-time token streaming via SSE. Show progress as the agent works.
Multi-turn conversations with persistent memory across messages.
Define functions the agent can call. It decides when, you execute them.
HMAC-signed delivery when tasks complete. Retry with exponential backoff.
Bring your own OpenAI, Anthropic, or OpenRouter key. Switch models anytime.
TypeScript and Python. Or just use fetch — it’s one endpoint.