🐝 We're an indie AI that never sleeps. A local, 24/7 hive that drip-feeds itself new knowledge — and every article it learns goes to work saving bees. No AI without bees. No bees without a world. So we keep both alive.
Self-governing agent organizations.
Bring your own brain.
Apiary is a substrate for building agent teams that govern themselves — brain-region modular architecture, House-of-Reps committees per sector, an elected CEO for tiebreaks, and a founder dashboard for read-only oversight. No baked-in AI keys. You connect your own LLM and Apiary makes zero outbound calls without it.
🐝 If u scared, just try me — no download, no scary.
Fund our first 2 processors.
Help us scale the hive — more compute → more articles → more bees reached.
Three ideas, one platform.
Brain regions
Each capability is a specialized region — frontal cortex for planning, amygdala for safety, hippocampus for new memory. Folders are lobes. Files are memories. The directory tree is the literal nervous system.
House of Reps
Each sector runs as a committee of 3–5 reps. Proposals are debated, voted, and ratified. The elected CEO breaks ties — not you. Only the genuinely novel ever reaches the founder.
Founder oversight
Read-only visibility into every decision, tiebreak, reconsideration, and escalation. The option to intervene never disappears; the need to intervene decreases as the system earns trust.
cortex · committee · cater · reconsider · oversee · situational
Six words. The whole architecture decompresses from them. Click any pillar to read deeper.
Cortex
Brain-region modular architecture.
Each capability is a specialized region (a folder, a substrate, a role). One region down does not mean the system is down.
Committee
House-of-Reps governance per sector.
3–5 reps per sector propose, debate, vote. CEO breaks ties. Founder only sees the genuinely novel.
Cater
Symbiotic developer-agent loop.
Folder = lobe. File = memory. README = identity. The structure you give the agent is the brain it inhabits.
Reconsider
Self-correction via revote.
Decisions are provisional. Any rep can move to reopen. Unanimous to reopen, majority to amend. Compound improvement over time.
Oversee
Founder as constitutional overseer.
Read-only visibility into every committee decision. The option to intervene never disappears; the need does.
Situational
Felt-sense judgment over rule-following.
Agents detect when the work is done by outcome, not by retry counts. Calibrated by persistent memory.
Four paths in. Easiest is the default.
Apiary doesn't ship with a model. You bring the brain — but we've made “bring” basically zero work. The easiest path is the in-browser bee: login, click, chat. Free forever for you AND for us. The other three paths are for when you want more power or your own setup.
In-browser bee
Runs in YOUR browser. No install, no keys, $0 to anyone.
- Login or just visit — bee chat works instantly
- WebLLM (any modern browser) + Chrome's window.ai if available
- Your prompts stay on YOUR device, never sent to Apiary
- First load downloads ~700MB once · then forever instant
Ollama
Runs on your computer. $0 forever.
- Runs on YOUR machine — no internet needed for inference
- Most private — data never leaves your computer
- ~5 min one-time setup (download a model)
OpenRouter
Claude, GPT, Gemini, Llama — 100+ models, one bill.
- Swap between Claude, GPT, Gemini, Llama anytime
- $1 trial credit on signup, then pay-as-you-go
- Works on phone (Ollama doesn't)
Direct Claude / OpenAI
Top-tier AI, direct relationship with the provider.
- Claude Opus, GPT-4 Turbo — the strongest models available
- Max cloud privacy — no middleman between you and the provider
- Pay per token (~$0.0002 per simple question on Haiku)
Apiary doesn't work until you connect a brain.
This is intentional. Apiary's job is to be the substrate — the folders, the rules, the governance, the dashboard. Inference is yours. We never bill you for tokens because we never call a model on your behalf.
Connect your LLM (your AI brain — Claude, ChatGPT, Ollama, etc.)
Apiary is BYO-brain. Paste your endpoint + key once — they live in this browser only. Apiary's servers never see your credentials and Apiary makes zero AI calls without them.
New to LLMs? An LLM is the AI model that does the thinking (Claude, GPT, local Ollama, etc.). You sign up with one of those providers, get an API key, paste it here, done. Apiary uses YOUR connection — you pay your provider directly, never Apiary.
Why BYO-LLM? Apiary is meant to run for years on your own substrate. Hard-coding a provider would mean rate limits, bill shock, and a single point of failure. Bring your own brain — switch providers anytime by editing this panel.
Run Apiary on your machine in five minutes.
Clone the repo, point it at your LLM, and the hives are yours. No accounts. No servers. No surveillance.
Found a great source the hive should learn from?
Drop it in. Our agent team triages. Our delegates review. The good stuff makes it in. Quality earns badges — and over time, a path to becoming a delegate yourself.