v0.6.0 release scope
These docs describe GraphForge v0.6.0. They replace the earlier user guides; there is no maintained pre-v1 documentation archive or backward-compatibility promise. Installation explains artifact availability and exact package versions. A source capability is not a claim that its release has already been published.
Start small
Section titled “Start small”The first experience is basic graph use: create connected data, ask a question, and inspect the answer. Choose an agent, a notebook, or an integration.
Additional capabilities are optional journeys. Users do not need the entire research model to use GraphForge.
| When you need it | v0.6.0 surface | Guide |
|---|---|---|
| Query and analyze connected data | openCypher, analyst verbs, Arrow results, graph construction | Basic graph |
| Keep and reuse work | Durable projects, reusable saved queries | Save and reopen |
| Remember what an inquiry established | Graph records for hypothesis, challenge, evidence, conclusion; optional native knowledge and provenance | Record an inquiry |
| Explore and review independent research | Slices, Branches, immutable Versions, comparisons, Proposals, Forks | Research workspaces |
| Evaluate externally produced decisions | Bounded context and validated choice, rubric, probability results; caller-owned policy | Decision workflows |
| Exchange research | Portable projects, Hub discovery, clone/publication, research history | Portable projects |
| Work at larger scale | Bounded ingest and interchange; workload-specific scale and benchmark qualification | Scale guidance |
Release readiness
Section titled “Release readiness”Draft release-note and announcement copy is in v0.6.0 notes (draft); it is not published.
M13 owns coordinated release readiness and publication across Rust, Python, Node, CLI, skills, and docs. #1096 tracks readiness; #1095 tracks publication. Use their live dependencies for status rather than inferring readiness from this feature list.
The included programs cover architecture maintenance, scale/interchange, benchmarks, the research lifecycle, and provider-neutral decision workflows. Research history and Hub command ergonomics continue through #1771 and #1772. Documentation must use the implemented command surface, not planned command names.
Guide readiness is tracked in #1208. #1209 qualifies clean candidate installs, real editor/agent/notebook paths, expected outputs, durable reopen, and independent first-use review. Technical examples do not establish nontechnical usability or adoption.
Core and associated applications
Section titled “Core and associated applications”Core owns graph and research semantics. The VS Code extension, XYG visualization, and website/Hub own their respective interfaces and hosting. A Core release does not qualify every consumer feature. The product map coordinates their work without changing implementation ownership.
Public projects support transparent participation and attributable contributions, using familiar Git author/committer practices. Private projects provide a non-public space for known collaborators. The hosting application enforces project access; local graph or governance metadata is not access control.
Beyond this release
Section titled “Beyond this release”Optional model-specific adapters, mobile bindings, and post-release community pilots have their own delivery work. They are not prerequisites for basic graph use. A browser-executed graph engine is not part of this release; a protocol validation module for Hub consumers is a separate capability.
Before v1.0.0, APIs and project formats may change without migration support. Current-version integrity, corruption refusal, recovery, and supported interchange remain required. A v1.0 compatibility policy will be a separate decision. See project compatibility.