Status
What exists, and how far it is verified.
This is a dated product snapshot, not a hosted uptime dashboard. Availability (the surface exists in source) and qualification (what has actually been verified) are separate columns of the same story.
Product source pinned at abc05bb48208· reviewed 2026-10-14
Chat
Chat entrypoint and TUI are present in the public tree as the daily interaction mode.
- Ordinary coding-loop reliability is under active qualification.
- Provider-backed sessions require local credentials and a working provider; quality varies by model.
Plan
Plan mode exists as a CLI/TUI entrypoint for planning before mutation.
- Planning behavior boundaries are documented in the product source.
- No live workflow qualification is claimed.
Deep
Deep mode exists as a higher-ceremony path intended for riskier work.
- Still maturing; route availability is not a reliability ranking.
- Deep is not a blanket reliability upgrade over Chat.
Doctor environment checks
babel doctor reports environment prerequisites (runtime, providers, Docker).
- Output is diagnostic; a clean doctor report is not a coding-reliability score.
Desktop harness
Windows Desktop bundles the real Babel CLI with provider credentials, enabled/default model selection, project files, approvals, persisted history and verified sandbox setup.
- Windows packages are unsigned; checksums verify bytes, not publisher identity.
- Automatic Desktop updates are unavailable.
- Active model-task shutdown on quit and cross-version profile migration remain unqualified.
- Plan and Deep currently require OpenRouter; other Desktop providers offer Chat.
Execution profiles
Profile policy (safe_repo default with Docker isolation, dev_local, and others) is implemented and covered by enforcement tests.
- The default safe_repo profile fail-closes without Docker.
- dev_local permits intrinsic inspection commands only; project-code tests/builds are denied without isolation.
MCP read-only inspection
A read-only MCP inspection entrypoint exists for control-plane style inspection.
- Inspection surface only; not a general execution API.
Ordinary coding-loop reliability
Everyday repository work reliability is explicitly under active qualification upstream.
- No general reliability claim is made or implied anywhere on this site.
- Scoped test passes are not a live-model quality score.
Distributions
What is published, and the evidence for it
Source checkout (clone + build)
Available · abc05bb48208 · signing: n/a
npm registry package (babel-harness)
Not available · signing: n/a
No npm registry package is published. Release v0.1.3 attaches a CLI tarball for reproducibility; the Windows installer bundles the CLI and Node runtime. Clone + build remains available for source users.
Windows x64 portable (Desktop + CLI)
Available · v0.1.3 · abc05bb48208 · signing: unsigned
Published asset · sha256:ebf2c6c928121a03c5d21364fbff636586ea80c7ce562ab14fe6b685ff752b9a
Windows x64 installer (Desktop + CLI)
Available · v0.1.3 · abc05bb48208 · signing: unsigned
Published asset · sha256:2cc94aee338edfa34059ad02aa508cd92062051741bff2eb8e7cec42b7aeffe1
GitHub release v0.1.0
Not available · v0.1.0 · signing: n/a
Release v0.1.0 (2026-07-23) is a source tag only — it carries no downloadable assets.
Qualification of the ordinary coding loop is an upstream engineering effort. When it produces new evidence, the pinned snapshot on this page is refreshed through a reviewed content change — never automatically. See how content is maintained.