Locus gives Codex and Claude Code a local memory layer for project structure, durable decisions, searchable history, and diagnostics. Built on MCP. Safe-by-default. Designed so your agent stops starting from zero.
Locus is already a real product line for Codex, not a future promise. The current release focuses on reliable memory primitives, diagnostics, and installable local workflows.
`memory_import_codex`, auto-import before `memory_search`, and Codex-aware diagnostics are already shipped in `v3.3.0`. Recent dialogue, prior decisions, and project structure can be recalled through one local memory system instead of ad-hoc summaries.
The shipped default remains `metadata`. That gives ingestion, visibility, and diagnostics first. Richer recall modes exist, but they are opt-in and should be treated with explicit privacy trade-offs.
`memory_status` and `memory_doctor` expose Codex-specific state so users can see what is imported, what is missing, and where the failure actually lives.
Repo-local Codex plugin packaging already ships. Marketplace-based one-command install and published runtime packaging are the next step, not hand-wavy marketing copy.
Most coding agents are strong inside a session and weak across sessions. Locus exists to bridge that gap without turning your workflow into a ritual of manual summaries.
Scan the codebase, map files and exports, and keep that structure searchable through MCP.
Save architectural choices and trade-offs so the next session starts with context, not guesswork.
Carbon Copy-style event ingest captures what happened, then makes it searchable through the same memory layer.
Diagnostics, audit visibility, and explicit capture modes matter as much as recall itself.
This landing page is intentionally simple. The real source of truth remains the repository: Quick Start, Codex setup, roadmap, releases, and implementation details all live there.