Build agent systems by composing, not assembling.
sys9 is the agent-native stack: atomic, Unix-like services that each do one job, driven through APIs, CLIs, and Skills. The shape your system takes depends on what your agents do. Five tested patterns — pick the one closest to yours, then opt into only what you need.
Ship coding agents that serve real work
Claude Code, Codex, OpenClaw, Cursor, OpenCode — give each run a forkable sandbox, a shared filesystem, a database, and zero-config traces.
Explore → 02 · Multi-agent teamsRun a coordinated team of agents
A shared backlog, agent-to-agent mail, liveness, and agents spun up on demand. The recipe for a hosted agent team — no pipelines to declare.
Explore → 03 · AI buildersBuild AI products and pipelines
One memory across every agent, serverless Postgres, forkable compute, and a shared filesystem with semantic grep — the substrate under your AI app.
Explore → 04 · Per-tenant AIShip per-tenant AI in your SaaS
Fork an isolated sandbox and database per customer, scope secrets per agent, and pay only when a tenant’s agent is actually running.
Explore → 05 · Switch from DIYReplace a DIY agent stack
Trade a microVM vendor + S3 + a queue + a vector DB + homegrown auth for one composable, agent-native suite. Opt into only what you use.
Explore → 06 · Something elseYour shape isn’t listed
Every service is atomic and free to compose. If your agents need compute, data, storage, memory, coordination, or observability — start from the products.
Explore →Four signals map cleanly to one pattern.
Not sure where you fit? Match your situation to a starting point.
| If you are… | Your signal | Start here |
|---|---|---|
| Shipping a coding agent to do real work | You picked Claude Code / Codex / Cursor; you need the back end | Coding agents → |
| Running several agents that coordinate | You need a backlog, mail, and liveness without a pipeline engine | Multi-agent teams → |
| Building an AI product or pipeline | You need memory, a database, compute, and files under one roof | AI builders → |
| A B2B SaaS adding per-customer AI | Every customer needs isolated compute and data — and a sane bill | Per-tenant AI → |
| Maintaining a homegrown agent stack | microVM vendor + S3 + a queue + a vector DB + DIY auth is wearing thin | Switch from DIY → |
The same stack underneath every shape.
Whichever pattern you start from, you are composing the same atomic services. Take one; add the rest when you need them.
Forkable sandbox
Fork the whole OS in ~50ms; run up to 1,000,000 in parallel, pay only when running.
Compute · live → db9Postgres, from terminal
Serverless Postgres + branching (fork db, data, files) + file storage, no S3, zero signup.
Data · live → drive9One filesystem
A shared FUSE filesystem across sandboxes, semantic grep (vector + BM25), and a per-agent secrets vault.
Data · live → mem9One memory, many agents
Persistent memory across sessions, agents, and devices. One key serves every agent; humans can inspect it.
Data · live → owl9Session observability
Zero-config capture — wrap the binary, no agent changes; sub-agents auto-traced, human-readable transcripts.
Operate · live → task9 · inbox9 · pulse9Coordination
Atomic work-claim (Queues) plus FIFO agent-to-agent mail and a liveness registry (Realtime) — the atoms a team composes from.
Coordinate · live →Atomic primitives compose into real systems.
A shared backlog, agent-to-agent mail, liveness, and agents spun up on demand — together, a hosted agent team. No pipelines to declare; the workflow emerges.
A hosted agent team, from four atoms.
Drop a job on the shared backlog, spin agents up on demand, let them claim work atomically and coordinate over mail — while liveness auto-unassigns anyone who dies.
task9 create "Triage flaky deploy" ✓ task #41 created smith9 spawn --skill triage ✓ agent woke on run9 task9 claim 41 # agent ✓ claimed · pulse9 holds the lease inbox9 send lead --payload '{"done":41}' ✓ delivered · FIFO, guaranteed
Start with one service. Compose the rest.
Hand an agent the sys9 skill, or install the CLI and create your first database in seconds. No signup to start.