Compare · Tier A
CodeHerder vs Stoneforge
An open-source (Apache-2.0) agent orchestrator: a Director plans, Workers execute in isolated worktrees, and Stewards test, merge, or hand back for review.
What Stoneforge actually ships
- A Director (strategic planner) routing to Workers (spawned per task, isolated git worktree each), reviewed by Stewards (automated test-run plus squash-merge on pass, or handoff-with-review-comments on fail)
- A background dispatch daemon and agent pools with per-type concurrency limits
- Multi-provider support (Claude Code default, OpenCode, or Codex), authenticating through the provider's own CLI with no API keys stored inside Stoneforge
- Event-sourced dual storage: SQLite as the cache, JSONL as the source of truth
Where it's strong
- Fully open source (Apache-2.0) with a clean, git-worktree-per-worker architecture close in spirit to CodeHerder's own
- Its stated problem, merge conflicts, duplicate work, and lack of visibility, reads as independent validation that this is a real category worth solving
How CodeHerder differs
- CodeHerder has moved further on cost and budget attribution, a device herd spanning multiple machines, cross-agent memory, human collaboration, and organizational (initiative → epic → story) structure, none of which Stoneforge publishes
Stoneforge is the closest OSS technical thesis to CodeHerder's own; the distinction is how far each has built past the shared starting point.
Capability snapshot
CodeHerder vs Stoneforge, capability by capability
Every row traces to the same source as the full matrix, where the full legend lives. In short: "not published" means no public evidence was found, not that the feature is absent. Verified as of 2026-08-08.
| Capability | CodeHerder | Stoneforge |
|---|---|---|
| Concurrent agents | Published | Published |
| Multi-vendor harnesses | Published | Published |
| Customer-owned machines | Published | Published |
| Multi-machine routing | Published | Not published (as of 2026-08-08) |
| Automatic work assignment | Published | Published |
| Capability-based routing | Published | Not published (as of 2026-08-08) |
| Atomic claiming | Published | Published |
| Dependencies | Published | Published |
| Custom workflows | Published | Partial (one fixed pipeline) |
| Server-enforced gates | Published | Published |
| Agent handoffs | Published | Published |
| Persistent shared memory | Published | Partial (JSONL state) |
| Live terminal / view | Published | Published |
| Human takeover / steer | Published | Partial |
| Cost per task | Published | Not published (as of 2026-08-08) |
| Cost per stage/model | Published | Not published (as of 2026-08-08) |
| Hard $ cap | Published | Not published (as of 2026-08-08) |
| Schedules | Published | Not published (as of 2026-08-08) |
| Webhooks / API / CLI | Published | Published |
| Self-host full platform | Published (Enterprise) | Published (OSS) |
Sources
Verified against Stoneforge's own site as of 2026-08-08.
- Stoneforge's homepage
- Stoneforge's core-concepts docs
Round up your herd.
Bring every human and every agent onto one table. Watch what's happening, see what's stuck, and know what it's costing you, live.
Already have a workspace? Sign in