Supported outcome

Find a plausible local model provider for a task and machine, qualify it through explicit checkpoints, and return a controlled loopback endpoint after exact approval.

Terminal result: LocalCode returns a verified provider capability and a separate record of why that provider was assigned; it does not run the coding task.

• tested-candidate · macOS on a 24 GiB Apple M4 Pro; Qwen2.5-Coder-7B-Instruct-3bit at revision a5ed5c3fb82a68cfd1682b910ac8a492249bbe3e; MLX-LM 0.31.3 and MLX 0.32.0 · Candidate supply selection, approved provider preparation, verified loopback readiness, no-restart reuse, exact status, disconnect, rollback and post-rollback restart · docs/releases/0.6.0-rc.1-acceptance.md
• unknown · Other Apple Silicon memory cohorts, Intel macOS, Linux and Windows · Candidate discovery, qualification, provider preparation and reuse · No installed 0.6.0rc1 acceptance run is retained for these environments

Non-goals

• Running or completing the user's coding task
• Claiming the best model for an arbitrary task
• Stable-release support or automatic model promotion
• Replacing an active provider without exact incumbent restoration
• Comparative energy optimization, remote-machine placement or long-running-agent supervision

Exact recovery boundary

Strategy: reinstall-exact-retained-predecessor

Disconnect: localcode disconnect

Prior retained version: 0.5.0-rc.1

State backup required: yes

• Disconnect or roll back active providers before changing product versions; active-provider migration is not supported.
• Product-version rollback requires the retained verified 0.5.0-rc.1 bundle and its compatible state backup.
• Cached or adopted model artifacts are retained across provider rollback and must be removed separately if no longer wanted.

Exact install command

uv tool install --force --constraints localcode-0.6.0-rc.1-runtime-constraints.txt ./localcode_prototype-0.6.0rc1-py3-none-any.whl