01
Memory-backed
Each ABot can carry continuity through AMS instead of starting from zero every session.
Automaton-Abots
Start with two custom agents that remember, share context, and keep working.
Automaton-Abots are custom agents with persistent AMS memory, shared knowledge, and a clean path into AOS governed runtime. Start with two custom ABots and 100 combined memories, then scale the fleet when the work is real.
Prime / 12 ABot Team Leads / 167 Specialist Agents
What An ABot Is
A custom worker with memory, role, policy, and history.
02
Role-specific
Engineering, growth, QA, PM, support, XR, and other specialists can be shaped for the work they own.
03
Prime-routed
Prime coordinates Team Leads and specialist ABots so work gets routed instead of scattered.
04
Runtime-ready
When the fleet needs schedules, trust tiers, containment, and audit trails, the same model grows into AOS.
Open Body · Interview-Born Brain
The body is open source. The brain is born by interview.
Open-source body
Fork the Abot runtime.
The Automaton-Abot body is the public runtime: channels, tools, sandbox, telemetry, MCP, AMS client, and security rails. Builders can inspect it, fork it, and run the body themselves.
Private brain
The interview creates the brain.
The interview creates the ABot brain: role, goals, voice, boundaries, channel behavior, memory scope, tool rules, and the working context it needs to become useful on day one.
Gateway targets
ABots meet you where the conversation already lives.
The gateway is designed around chat adapters, not another isolated inbox. The open-source body currently names Telegram, Discord, and Slack adapters, with WhatsApp configured as a Baileys-compatible target. AMS dashboard chat carries native web sessions back into the same agent runtime.
Discord
Gateway events and channel messages for communities, squads, and ops rooms.
Slack
Socket Mode and chat posting for internal teams and workspaces.
Telegram
Bot API, webhook, or polling flow for lightweight direct chat.
WhatsApp
Baileys-compatible gateway target for business and field communication.
AMS Chat
Native dashboard sessions tied back to ABot synthesis and memory.
AMS Free Tier
Build the first pair of custom ABots for free.
Open for business
$0/ month
AMS Free is the starting lane for Automaton-Abots. It gives one user enough memory and structure to build, test, and keep two custom ABots useful.
- 2 custom Automaton-Abots
- 100 AMS memories combined across both ABots
- Shared memory and handoffs across the first two ABots
- H-MEM structure for episodic, semantic, and procedural memory
- Prime routing model for Team Lead and specialist organization
- Upgrade path into larger AMS memory and AOS governed runtime
Scale Path
Start with two. Grow into a fleet.
Free
Two ABots
Prototype your first custom agents with 100 combined memories and a clean Prime and Team Lead organization model.
AMS Professional
More memory
Scale into larger memory budgets, memory linking, procedural memory, automata generation, and production support.
Automaton OS
Governed runtime
Add schedules, trust tiers, containment triggers, MCP mesh, telemetry, and marketplace packaging for the full fleet.
Prime And AOS
Prime routes the fleet. AOS governs execution.
Prime
Prime is the orchestration layer: it sees goals, chooses the right Team Lead, routes work to specialist ABots, and writes outcomes back into AMS.
AOS
AOS is the governed runtime: schedules, policy cards, trust tiers, containment, telemetry, and installable skills for agents that need to run in the open.
Next Step
Use Free for the first two ABots. Scale when the fleet starts working.
The free tier is the front door. When the ABots need larger memory, more automata, shared governance, or a production runtime, move into AMS Professional and Automaton OS.