Stackon
Join waitlist
Private alpha · observability-first

Ship faster.
With receipts.

Stackon is the agentic developmental environment for engineering teams who refuse to fly blind. Multi-agent orchestration, BYO compute, and a creator marketplace — instrumented end to end.

Start shippingJoin Discord

● Recently shipped · v0.39.4

Press-kit full lockup — mark + wordmark on one canvas

trace · speciment+00:00:14.2
trace · run_8c4f742ms · $0.0045
agent.plan
742ms
tools.search
118ms
llm.completion
284ms
tools.edit
92ms

eval · gate.passed

PR check ready to merge

5/5 assertions$0.0045 · 0 retries
cost · this monthteam · acme

$312 / $1,000

31% · ↘ 18% wow

canvas$184
evals$93
adversary$35
5 spans · 1 eval · 1 budget meterlive capture

3,200/min

Spans ingested

Streaming live · per team

p95 <500ms

Trace ingest latency

Postgres + JSONB · ClickHouse on deck

18/20

Pillars shipped

Phase 2 · 3 · 4 — auto-counted

Built to ship together

Four products.
One workflow.

Everything an engineering team needs to go from intent to production — with agent teammates at every step of the ship loop.

Trace.

See every span. Replay any run. Ship with receipts.

trace · run_8c4f742ms · $0.0045
agent.plan
700ms
tools.search
154ms
llm.completion
266ms
tools.edit
126ms

Canvas.

Wire a Planner → Coder → Reviewer pipeline. Watch them debate.

Planner
Coder
Reviewer

MCP.

One JSON block. Every tool wakes up connected to your workspace.

{
  "mcpServers": {
    "stackon": {
      "command": "npx",
      "args": ["@hunterade/mcp"]
    }
  }
}

Mobile.

Approve queued runs from your phone. Watch traces land in real time.

Approvals3 pending
PR review pipeline3 nodes · 1m ago
→
Schema migration1 node · 4m ago
→
Release notes2 nodes · 9m ago
→
§ 02Lead pillar
Trace · shipping

Every run
leaves a trace.

Typed spans for every LLM call, tool, and child task — cost and latency captured by default. Replay any run, days later. Prove this week beat last week.

Trace · run_8c4ft+00:00:14.2 · live capture
latency742mscost $0.0045
agent.plan
742ms
$0.0024
tools.search
118ms
$0.0003
llm.completion
284ms
$0.0017
tools.edit
92ms
$0.0001
evals.assert
61ms
—
eval · pass5/5 spans · 0 retries

otel · jsonb · pgvectorspecimen / 005

01

OTel-compatible

Bring your existing observability stack along — no rewrites, no proprietary schema.

02

Cost on every row

Token spend rolled up per span, per task, per team. Budgets and alerts shipped in.

03

Eval gates

PRs blocked the moment a trace regresses against last week. Receipts attached.

Get on the alpha

The agentic stack

From Prompt.
To Production.

Describe the task. Orchestrate a team of agents. Ship a measurable result. Stackon is the loop that closes itself — every run leaves a trace.

01 · Describe

Describe.

Open a Mission. Drop a task description in 3-6 lines and let the Planner agent decompose it into concrete subtasks across 30+ languages.

mission · M-0042● draft

Ship fresh-token rotation to /api/v1/auth.

// Acceptance: existing access tokens stay valid until expiry; new column on api_tokens for hashed refresh secret; migration ships in same PR.

Planner ready · 4 subtasks proposed

02 · Orchestrate

Orchestrate.

Wire a canvas: Planner → Coder → Reviewer → Tester. Watch them debate. Stackon hands the baton between agents that already share your repo, ADRs, and prior traces.

canvas · auth-rotation● running
Planner

claude-opus-4-7

Coder

claude-opus-4-7

Reviewer

claude-opus-4-7

Tester

claude-opus-4-7

03 · Ship

Ship.

Evals run on every PR. Budgets catch the runaways. Approvals fan out to mobile. The same trace that proved this run worked will prove next week's didn't regress.

eval · gate.passed5/5 assertions

✓ ship/auth-token-rotation

PR #1142 opened · 1 review requested

$ stackon trace receipts

→ 16 spans · 742ms · $0.0045 · 0 retries

§ 03Ecosystem

The full suite.
Built to ship together.

The platform is a constellation of products, each fluent in the same spans, the same identity, the same budgets. Build the slice you need first.

obs · receipts

Observability

Prove every agent run. Traces, evals, postmortems, adversarial sweeps, cost — instrumented end to end.

build · orchestrate

Canvas & Missions

Compose multi-agent pipelines visually, version them like code, ground them in your repo and ADRs.

everywhere · trusted

Apps & Surfaces

Reach your agents from any surface. CLI, MCP, phone, voice, BYO compute — same trace, same identity, same trust.

  • Claude Code
  • Cursor
  • Windsurf
  • Codex
  • Zed
  • Stackon MCP
  • Stackon CLI
  • VS Code
  • JetBrains

The loop

Run, trace, improve. Then run again — with measurable confidence.

01

Run

Spin up agents from the canvas, MCP, or HunterCLI. Single shot or coordinated multi-agent.

02

Trace

Every LLM call, tool, and handoff is captured as a typed span with cost, latency, and IO.

03

Improve

Evals run on every PR. Budgets and red-team attacks catch regressions before they ship.

§ 04Receipts

We ship in public.
Read along.

Full changelog
shipped2026-05-22

Press-kit full lockup — mark + wordmark on one canvas

Added a horizontal lockup to the press kit — mark on the left, wordmark on the right, both centered on the same 1536×1024 canvas. Ships in both dark (`#0A0A0A`) and light (`#FAFAF8`) variants. Useful for footers, slide decks, and any context where a journalist or partner wants both identity pieces composed as a single asset rather than reassembling the mark and wordmark by hand.

v0.39.4

fix2026-05-20

Press-kit wordmark — designed SVG replaces gpt-image-1 raster

The `/media` wordmark slots (`Wordmark · dark` and `Wordmark · light`) were generated by gpt-image-1, which mangles text — the rendered `STACKON` letterforms had subtle distortion and paper-grain artifacts that didn't read at press scale. Replaced with a hand-wired vector wordmark built from Geist Regular glyph paths.

v0.39.3

preview2026-05-20

Stackon for Windows — Tauri shell scaffolded (NSIS + MSI)

The Tauri shell now targets Windows alongside macOS. `tauri.conf.json:bundle.targets` extends to `["dmg", "app", "nsis", "msi"]` — Tauri silently skips the platform-foreign targets at build time, so the same `npm run desktop:build` invocation produces a DMG on macOS and an NSIS .exe + MSI on Windows. No code changes in `main.rs`; the Rust shell is platform-agnostic by design.

v0.39.0

Team shipping moment
The team, watching a trace land.
§ 05Apply

Be among the first teams to ship agents you can trust.

Drop your email. We’ll send your seat the moment we open another wave — small batches, high touch.

No spam · Unsubscribe in one click