Public beta · ships Q3 2026

Run your Claude Code agents on a schedule.
Watch every run.

FerrFleet is a thin scheduler, catalog, and observability layer on top of the Claude Code CLI. Catalog your agents, trigger them on demand or on a cron, tail the live transcript, and track tokens and cost per run.

Live · run_8x9k2j · flow-release-bot
→ trigger: cron (0 9 * * 1)
→ claude -p "draft the release notes…"
→ tokens in: 4.1k
[assistant] Reading merged PRs since v3.1.0…
[tool_use] Bash: git log --oneline
[assistant] Drafting CHANGELOG entries…
→ tokens out: 1.9k
✓ done in 38s · cost $0.082
Spend · last 14 days
RUNS
557
COMPUTE COST
$12.40
flow-release-bot
31 runs
track-triager
412 runs
vault-rotator
96 runs
growth-seo-auditor
18 runs
What FerrFleet does.

A control plane for
your Claude Code agents.

We don't reimplement the model, tool use, or context management — Claude Code does that. FerrFleet catalogs your agents, runs them, and shows you what happened.

01

Agent catalog

Version your Claude Code agents in one place: prompt, working directory, sub-agents, and the connectors they can reach. Edit, fork, and run from the dashboard.

02

Triggers

Run an agent on demand, on a cron schedule, or from a signed webhook. Each schedule belongs to an org and fires a fresh run.

03

Cost and token observability

Every run records tokens in, tokens out, and the compute cost we derive from them, attributed to the agent that ran. See spend by day and by agent.

04

Live transcript

Tail a run as it happens over SSE: prompts, tool calls, and assistant output stream into the dashboard. Re-open any past run to read the full transcript.

05

Connectors

Give an agent typed access to MCP servers and the rest of the FerrLabs suite. Connector secrets are resolved per run and never stored in the agent definition.

06

First-party fleet

A catalogue of agents that drive the FerrLabs suite — vault-rotator, track-triager, flow-release-bot, growth-seo-auditor. Free to run when you already pay for the target product.

Spend you can see

Know what each agent actually costs.

Every run records tokens in and out and the compute cost we derive from them. The Spend page rolls that up by day and by agent — no guesswork at month end.

Cost per agent · last 14 days 4 agents tracked
flow-release-bot
31 runs
$0.082
track-triager
412 runs
$0.002
vault-rotator
96 runs
$0.018
growth-seo-auditor
18 runs
$0.124
How it fits together

Built on Claude Code.

Engine
  • Claude Code CLI
Triggers
  • Manual
  • Cron schedule
  • Signed webhook
FerrLabs suite
  • FerrVault
  • FerrTrack
  • FerrFlow
  • FerrGrowth
Runtime
  • Kubernetes (Job per run)

Ship agents that
survive Monday morning.

Dispatch, retry, observe, bill — one platform.

Request access →