Frequently asked questions
What KodeAtlas does now—and what it is being built to do next.
Current archival capabilities and future intelligence layers remain deliberately separated in every answer.
What is KodeAtlas today?
KodeAtlas CLI v0.2.0 is a deterministic, resumable, verifiable Git and GitHub archival CLI. It preserves source history and collaboration evidence together with provenance, indexes, completeness declarations, and SHA-256 integrity manifests.
What is KodeAtlas becoming?
The product direction adds temporal TypeScript/JavaScript symbol history, durable code anchors, context packets, backlog staleness, regression memory, decision memory, and multi-agent coordination in explicit phases.
Is the context and coordination platform available now?
No. The website labels those capabilities in development or planned. The archive foundation is the current product baseline.
Is KodeAtlas another GitHub backup tool?
Archival is foundational, but the long-term differentiation is the intelligence layer above collection: semantic identity through time, cited context, planned-work synchronization, risk memory, and agent coordination.
Does KodeAtlas replace GitHub, Jira, Linear, CI, documentation, or coding agents?
No. It is designed to connect and operationalize their evidence at the code boundary where a developer or agent needs to act.
Why are line numbers not enough?
A permanent link is useful historical evidence, but it intentionally stays bound to one commit. A living task or decision needs a semantic anchor that can resolve through moves, renames, signature changes, extractions, splits, merges, or deletion.
Will KodeAtlas claim that a commit caused a regression?
Only when the evidence satisfies a defined confirmed threshold. Other relationships are labeled strongly associated, possibly associated, or temporally related, with rationale and source links.
Will KodeAtlas use embeddings?
Possibly for fuzzy narrative retrieval and historical search, but compiler, AST, Git, and structured graph evidence define identity and relationships.
Why start with TypeScript and JavaScript?
The product needs a narrow vertical with compiler-accurate symbols and a strong evaluation corpus before it expands across languages.
How will KodeAtlas coordinate agents?
The planned model uses authenticated, expiring semantic claims over symbols, dependencies, tests, schemas, endpoints, contracts, branches, and worktrees rather than locking entire files by default.
Can it run locally?
Yes. Local-first is the recommended default for the first context workflows so teams retain control of repositories and archives without mandatory hosted code transfer.
Who is a good early-access partner?
A technically engaged TypeScript/JavaScript team with meaningful Git/GitHub history, distributed rationale, stale-task or regression-memory problems, and willingness to evaluate whether cited context changes agent plans.
Take the next step
Have a repository where the missing context is measurable?
Start with the archive foundation or bring a real TypeScript/JavaScript context problem for the next product phase.