Build your own AI agent, grounded in your own documents
Upload documents or crawl URLs into a knowledge base, wire up a retrieval-augmented agent on a visual canvas, and chat or integrate it via API — all with your own LLM provider keys.
Everything to build one agent
Not a suite of products bolted together — one focused pipeline from raw documents to a working, integrable agent.
Knowledge base ingestion
Upload PDFs or crawl URLs. Documents are parsed, chunked, and embedded automatically so your agent can retrieve exactly the right context.
Visual workflow canvas
Wire your Knowledge Base, Agent Settings, and Model Key together on a drag-and-drop node canvas — see the whole pipeline at a glance.
Bring your own LLM keys
Use the platform key to get started free, or plug in your own OpenAI, Claude, Gemini, DeepSeek, or Groq key with automatic fallback ordering.
Tunable agent settings
Control system prompt, temperature, and input/output token limits so responses match exactly how your agent should behave.
Live chat playground
Test your configured agent in a real conversation before you ship it, with retrieval context pulled straight from your knowledge base.
Programmatic API access
Generate an API key and call your agent from cURL, Python, or TypeScript — the same chat pipeline you tested in the playground.
Ingest
Upload documents or crawl a URL — Sabai chunks and embeds the content into your personal knowledge base.
Configure
Wire up the workflow canvas: pick your model key, set the system prompt, and tune retrieval and generation.
Ship
Test live in the chat playground, then call the same agent from your own app via a simple API key.
Ready to build your agent?
Create a free account and go from a blank knowledge base to a working agent in minutes.
Create your account