← All implementations

WorkOS · Platform

Project Horizon

An internal autonomous 'code factory' where a continuously running swarm of agents handles the implementation loop while engineers focus on requirements and acceptance testing. Deliberately modular so the harness can evolve.

1 Supports2 Supports3 Contextualizes

Supporting infrastructure

This entry describes supporting infrastructure that other work builds on. The catalog classifies it as a platform. The record also reports a workflow.

Approach type
Platform
Work
Coding, Code review, Security
Human involvement
Drafts reviewed
Invocation
Event driven, Interactive
Interfaces
Linear, Github, Slack, Web
Deployment stage
Deployed
Evidence strength
Detailed primary
Entry reviewed

How it works

The workflow the sources report for this implementation.

Separate what runs code from what manages the lifecycle

1

A runtime controlled end-to-end, with lifecycle APIs and egress controls for the threat model

1

Each run ships work and produces the next set of fixes, surfacing where the platform is brittle

1

Tuning tools is ongoing, not a one-time integration

2

Where people stay involved

  • requirements and acceptance criteria → tested implementationOutcome review · Level 4

Level 4 for requirements and acceptance criteria → tested implementation; human attention boundary: outcome-review.

1

Observation date
2026-05-06

Implementation details

Sandbox

Cloudflare Containers + Sandbox SDK; disposable, tightly scoped sandboxes with explicit lifecycle APIs and egress controls; full monorepo stack in Docker dev containers

1

Harness

Modular by design; the core article runs OpenCode in the sandbox; the Applied AI Showcase runs Claude Remote Routines. The harness is swappable as agent tech changes; separate PM, implementation, and prospective verification/security roles

1 Supports2 Supports

Interfaces

linear, github, slack, web

1

Tool access

A custom MCP server stitches internal data sources (Datadog, Sentry, Slack, WorkOS Pipes); all outbound traffic proxied through Workers with allowlists, limits, logging, and token injection

1 Supports2 Supports

Knowledge

AGENTS.md and CLAUDE.md capture scripts, docs, conventions; MCP codifies the patterns engineers already follow; Notion + Figma for specs/mockups

1

Credentials

WorkOS Pipes (no OAuth/token-refresh to maintain); scoped short-lived GitHub tokens per user; engineers use their own identity in the MCP; least-privilege + egress controls

1

Context management

The orchestrator pauses/resumes sandboxes and tracks state + artifacts across a run

1

Reported results and limitations

The catalog records what the sources report, with the scope and the denominator of every figure. A qualification below limits the figure it sits under.

Lessons and interpretation

You need purpose-built agent infrastructure; a runtime you control end-to-end with lifecycle APIs and egress controls

1

Separate concerns: sandboxes are an execution primitive; the orchestrator is the control plane

1

Build modularly so the harness can evolve; OpenCode today, Claude Remote Routines tomorrow, without rebuilding the platform

1 Supports2 Supports

Make autonomy a platform; the system gets faster and more reliable through use as fixes feed back in

1

MCP tuning is an iterative product, not a one-time integration; codify the patterns engineers already follow

2

Sources and research details

Citations link to the original publisher. Each source also keeps a preserved copy in the repository, so a changed or removed page stays checkable.

  1. Project Horizon - an autonomous code factory at WorkOShttps://workos.com/blog/project-horizonEngineering blog · First party · Last source verification: 2026-08-31
    Preserved copy ↗
  2. Applied AI Showcase (Horizon with Claude Remote Routines)https://workos.com/blog/applied-ai-showcaseEngineering blog · First party · Last source verification: 2026-08-31
    Preserved copy ↗
  3. An autonomous UI-quality programhttps://workos.com/blog/autonomous-ui-quality-programEngineering blog · First party · Last source verification: 2026-08-31
    Preserved copy ↗
  4. Hacker News submission for Project Horizonhttps://news.ycombinator.com/item?id=48039227Hn thread · Community · Last source verification: 2026-08-31
    Preserved copy ↗
Research details for every claim on this page
  1. Summary
    Statement type
    Fact
    Provenance
    Reported
    Confidence
    High
    Confidence reason
    A linked first-party source states the claim.
  2. Sandbox
    Statement type
    Fact
    Provenance
    Reported
    Confidence
    High
    Confidence reason
    A linked first-party source states the claim.
  3. Harness
    Statement type
    Fact
    Provenance
    Reported
    Confidence
    High
    Confidence reason
    A linked first-party source states the claim.
  4. Model
    Statement type
    Fact
    Provenance
    Reported
    Confidence
    High
    Confidence reason
    A linked first-party source states the claim.
  5. Interfaces
    Statement type
    Fact
    Provenance
    Reported
    Confidence
    High
    Confidence reason
    A linked first-party source states the claim.
  6. Tool access
    Statement type
    Fact
    Provenance
    Reported
    Confidence
    High
    Confidence reason
    A linked first-party source states the claim.
  7. Knowledge
    Statement type
    Fact
    Provenance
    Reported
    Confidence
    High
    Confidence reason
    A linked first-party source states the claim.
  8. Credentials
    Statement type
    Fact
    Provenance
    Reported
    Confidence
    High
    Confidence reason
    A linked first-party source states the claim.
  9. Context management
    Statement type
    Fact
    Provenance
    Reported
    Confidence
    High
    Confidence reason
    A linked first-party source states the claim.
  10. Supporting component
    Statement type
    Fact
    Provenance
    Reported
    Confidence
    High
    Confidence reason
    A linked first-party source states the claim.
  11. Supporting component
    Statement type
    Fact
    Provenance
    Reported
    Confidence
    High
    Confidence reason
    A linked first-party source states the claim.
  12. Supporting component
    Statement type
    Fact
    Provenance
    Reported
    Confidence
    High
    Confidence reason
    A linked first-party source states the claim.
  13. Supporting component
    Statement type
    Fact
    Provenance
    Reported
    Confidence
    High
    Confidence reason
    A linked first-party source states the claim.
  14. Lesson
    Statement type
    Inference
    Provenance
    Catalog judgment
    Confidence
    Medium
    Confidence reason
    The catalog derives this observation from the linked sources.
  15. Lesson
    Statement type
    Inference
    Provenance
    Catalog judgment
    Confidence
    Medium
    Confidence reason
    The catalog derives this observation from the linked sources.
  16. Lesson
    Statement type
    Inference
    Provenance
    Catalog judgment
    Confidence
    Medium
    Confidence reason
    The catalog derives this observation from the linked sources.
  17. Lesson
    Statement type
    Inference
    Provenance
    Catalog judgment
    Confidence
    Medium
    Confidence reason
    The catalog derives this observation from the linked sources.
  18. Lesson
    Statement type
    Inference
    Provenance
    Catalog judgment
    Confidence
    Medium
    Confidence reason
    The catalog derives this observation from the linked sources.
  19. Operating model assessment
    Statement type
    Inference
    Provenance
    Catalog judgment
    Confidence
    High
    Confidence reason
    The source describes engineers focusing on requirements and acceptance testing while agents run the implementation loop.
    Observation date
    2026-05-06

The catalog records no related implementation for this entry yet.

Back to the directory →