An AI that actually
knows your world.
Lifebot lives inside your graph — the AI behind your chats, your search and your automation. Ask it in any node and it answers in context, acts with your say-so, and runs on the model you pick.
More than a chatbot.
Answers in context
Ask inside any node and Lifebot answers knowing that node — its chat, children and the scope you allow.
Searches your graph
“Where did we decide the pricing?” — it finds the node and the message, across everything you’ve shared with it.
Creates & edits nodes
Tell it what you need and it adds tasks, builds a whole sub-project in one go, or drops a reference link — with your say-so.
Runs on its own
Wired into a contract, Lifebot acts on a trigger — summarizing, drafting, watching — not just replying.
Stays silent until asked
It won’t butt into every chat. Mention it, or turn it off per message — you decide when it speaks.
Respects your secrets
Wallet nodes are invisible to it. Lifebot never sees what you’ve marked private to yourself.
Four providers. A fast one and a smart one each.
Lifebot isn’t tied to one vendor. Set a default for the whole app, override it on any node — and flip intelligence on where it matters: a fast, cheap model for everyday automation, the flagship model for the hard calls.
Claude
A fast Claude for the everyday, the flagship for the hard calls.
OpenAI
A quick GPT for volume, the top model for depth.
Gemini
A light Gemini for speed, the deep one when it counts.
DeepSeek
An open, low-cost pair — one fast, one deep.
Four providers, each with a fast and a smart tier — and any new OpenAI-compatible model (Kimi, Groq, Together…) drops in with a single line. A little intelligence indicator shows which tier a node is on.
Pick its brain — and its reach.
On any node you can override the app-wide default: choose the provider (Claude, OpenAI, Gemini, DeepSeek — or any OpenAI-compatible model), turn on a smarter model where it matters and leave the rest light and cheap, and set how far it may look — just this node, out across your graph, or never into your private data. For a shared or sensitive node, keep it to exactly what’s inside.
Per person, not per node. Because Lifebot runs on the token of whoever asked, each teammate sets their own model and reach — and it reasons over their tree and replies to them. Your settings and your data never cross into someone else’s.
It writes your automation — and fixes it.
Lifebot knows how contracts work — the events, the injected variables, the syntax. Describe what you want and it builds the whole thing: the contract node, a reference to the data to watch, and the code (or a plain-language prompt) in a node it creates. When that code breaks, it reads the error, iterates, and switches the contract back on.
Builds the contract
Code or prompt — Lifebot creates the contract node, the code node, and wires the config, from a single sentence.
Points it at your data
It drops a reference to the exact nodes you want watched or analyzed, so their events flow up and trigger the contract.
Iterates on the code
Reads the error posted to the code node, rewrites the code, verifies, and re-enables it — a self-healing loop.
Knows the syntax
Ask it anything about contracts — which events fire, what’s injected, how to call the SDK — it has the docs by heart.
Your second brain, awake.
It remembers, it searches, it acts — in your context.
Meet Lifebot — free →