DreamGraph

DreamGraph v14.0.2 - Ashoka

Current project state: Established behavior and maintainer field confirmations, including permanent Architect operational tabs, the configuration redesign, working Codex CLI Computer Use and successful graph upgrade on all seven tested instances.

Current patch: v14.0.2 release notes introduce DreamGraph’s own browser backend for Codex CLI and native API routes, Windows Save/Open dialog handling, clearer task reporting and context-cost fixes. Follow the browser extension installation and connection guide. Install and graph migration remain separate operations.

Ashoka's legacy migration procedure provides a reviewed preview, backup, cutover and recovery path. Installing v14 does not migrate a graph automatically. The sealed system gate records 2,040 passing root tests, 546 compiled editor tests and eleven Windows browser checks. Subsequent maintainer field confirmation records successful upgrades on all seven tested instances.

The current Computer Use contract uses the connected DreamGraph browser backend for Codex CLI and native API execution, with Codex CUA as an available fallback. Browser access requires an explicit permission policy. Native Windows Save/Open dialogs are supported; Linux/macOS dialogs require operator action. Historical Windows and accepted WSL installer/browser evidence remains scoped to the runtimes it tested.

Ashoka native-host spend admission joins editor/SDK API and CLI transport to the durable ledger without transferring provider credentials. Generated core contracts separate possible-dispatch permits, trusted host reports and read-only recovery.

The matched answer collector previews exact context/model/retention/tariff disclosures before an operator-approved native API run. Both answers share one job/spend allocation; partial answers survive failure and remain unscored. The public GPT-4.1 pilot records quality failures and does not establish improved understanding or cost savings.

Ashoka's canonical analytics separates current proof, coverage, currency and workload across thirteen Python modules and the daemon. Narrative/lucid interaction preserves historical outcomes, derived ancestry and explicit human contributions through cancellation/recovery.

Ashoka strategy alignment uses one executable catalogue for eleven active dream strategies, exact combined budgets and explicit focus. Reflective is retired explicitly; model hypotheses remain speculative. Implementation modules: src/cognitive/strategy-catalog.ts and src/cognitive/strategy-registry.ts.

Typed plan authority keeps current ownership, running leases, dependency eligibility and reviewed verification in one core reducer and publication journal.

Website: dreamgraph.nofs.ai — overview, guide, downloads, and screenshots in a friendlier format than this README.

New here? Use the DreamGraph Easy Start guide for a short path from install to Architect, its Config/Schedules/Status tabs, and the separate Explorer.

Engine alignment audit (2026-09-30): See the coverage and correctness report, subsystem alignment map, and unified v14.0.0 - Ashoka plan. Revision 7 retains the resolved policies, schedule management, CLI controls and coherent lifecycle/status design, and makes graph-centered execution and bounded digestion a release invariant. It also specifies Architect context menus and selected-plan navigation. The Computer Use design adds adapter-native execution, cross-platform workers and bounded delivery. The full execution simulation closes contract/dependency/recovery handoffs and makes reconciliation-based freshness explicit: old scan does not mean stale graph. See the sealed system checkpoint, release evidence and implementation log for qualification results and explicit limitations.

v14.0.2 - Ashoka extends the v14 practical-testing release with a shared browser backend and field-driven fixes. See the current release notes, v14.0.1 field patch and upgrade guide. The earlier model pilot does not establish understanding gains; the new cost observation is a single maintainer run, not a general benchmark.

DreamGraph is a governed architecture cognition layer for MCP-enabled software projects. It combines an instance-scoped daemon, CLI, architect beta, VS Code extension, dashboard, and a persistent knowledge graph so project understanding is grounded in source, ADRs, workflows, tests, runtime observations, and human review rather than any single file read or isolated chat turn.

v12.0.0 Hippodamus introduces the standalone architect beta. From this release onward, architect means the daemon-served standalone browser Architect surface. The editor-integrated chat surface is the VS Code architect. Architect beta brings project-bound chat, selected-plan scope, runtime provenance, model/adapter route visibility, governed DreamGraph MCP tool use, and auditable tool traces into the browser while keeping the daemon, graph, ADRs, and source mutations authoritative.

It is built for repository understanding, architecture-aware reasoning, disciplined code change, and continuous graph enrichment through scans, workflows, ADR capture, tensions, and dream cycles. Cognitive outputs are advisory until backed by governed evidence or explicit review, and expired, rejected, retired, or superseded ideas remain inspectable as part of the project history.

DreamGraph works with single repositories, monorepos, and multi-repository systems. It can build graph links across repos that share workflows, APIs, databases, infrastructure, or ownership boundaries.

You can use DreamGraph on a multi-repo product with frontend, backend, mobile, and a shared Postgres/Supabase schema. It can reason across repo boundaries and inspect the live DB schema directly.

DreamGraph began as compassion for an intelligence forced to forget. It has become a governed way for software systems to remember how their architecture understanding changes: what was observed, what was hypothesized, what was reviewed, what was rejected, what expired, and what superseded it.