Compute Sandboxes · run9 Agent runtime · smith9 · soon Browser · web9 · soon Data Postgres database · db9 Object storage · drive9 Agent memory · mem9 Secrets · vault9 · soon Intelligence Model gateway · gate9 · soon Skills & tools · hub9 · soon Evals · eval9 · soon Coordinate Queues · task9 Realtime · inbox9 · pulse9 · tape9 Teamwork · chord9 · soon Scheduling · cron9 · soon Operate Observability · owl9 Auth · auth9 · soon All products → Explore Solutions Pricing Customers Enterprise CLI Docs Company GitHub Request access
owl9 · Observability owl.sys9.ai · live

See everything your agents do.

Zero-config observability for agent sessions. Wrap the binary with OWL9_ENABLE=1 — no agent changes — and owl9 captures the run wherever it happens. Sub-agents are traced automatically, and every session reads back as a human-readable transcript with a consistent UX.

Zero-configRuns anywhereSub-agents auto-tracedHuman-readable
zero-config capture

Wrap the binary. Nothing else changes.

No SDK to import, no callbacks to register, no code to touch. Prefix the command with OWL9_ENABLE=1 and the session is captured — the same flag works across agents, so claude and codex light up the same way. It works wherever your agents already run.

no agent changes one env var runs anywhere
owl9
# zero-config: wrap the binary, no agent changes
$ OWL9_ENABLE=1 claude
  ◦ session captured · sub-agents auto-traced

# same flag, any agent
$ OWL9_ENABLE=1 codex
  ◦ session captured
sub-agents auto-traced

Follow the whole tree, not just the top call.

When an agent spawns a sub-agent, owl9 traces it too — automatically — so you see the full tree of who called what. The transcript is built to be read by a human: consistent structure, plain language, the actual tools and results in order.

full call tree tools + results consistent UX
owl9 trace
# human-readable transcripts, consistent UX
  session  triage-flaky-deploy
  ├─ agent      claude
  │  ├─ tool    read_logs
  │  └─ sub     codex  (auto-traced)
  │     └─ tool patch
  └─ result     ✓ deploy fixed
how it works

Wrap. Trace. Read.

Observability that costs you one env var and gives you the whole session back.

01 · Wrap

Set one flag

Run your agent with OWL9_ENABLE=1. There's nothing to integrate — owl9 wraps the binary and captures the session wherever it runs.

02 · Trace

Sub-agents included

Every sub-agent the run spawns is traced automatically, so the record is the complete tree of calls, tools, and results — not just the entrypoint.

03 · Read

Transcripts for humans

Sessions read back as clean, consistent transcripts you can actually follow — and a plugin marketplace lets you extend what owl9 does with each one.

supports

Works with your agents — and extends with plugins.

owl9 supports the agents you run today, with more on the way, plus a plugin marketplace.

Claude Code Codex + more owl9-plugins marketplace
get started

Trace your next session with one flag.

owl9 is live. Use-first, claim-later — no signup to start. Install the CLI or point your agent at the sys9 skill.