We ship in public.
Every meaningful change to Logic64 is documented here — features, improvements, fixes, and infrastructure updates.
Streaming generation + SSE output
The generation engine now streams progress over Server-Sent Events. Watch your workspace materialize in real time — file by file, phase by phase.
- SSE endpoint for live generation progress (phase, current file, % complete)
- BullMQ job queue redesign — jobs now survive engine restarts without requeuing
- Dashboard stream panel shows live token output with scrollback
- ACK endpoint separated from stream — no inline acknowledgement over SSE
- Concurrency limit per user enforced at queue level, not application level
CLI: logic64 pull
The Logic64 CLI is live. Authenticate once, pull any of your generated workspaces directly to your local filesystem.
- `logic64 login` — OAuth device flow, stores token in OS keychain
- `logic64 pull <workspace-id>` — downloads and extracts workspace to current directory
- `logic64 ls` — lists all generated workspaces for your account
- Atomic writes — partial downloads never leave a corrupted workspace
- Published to npm as `logic64` — works on macOS, Linux, and Windows
Multi-project support + project settings
Accounts now support multiple projects. Each project has its own architecture spec, generation history, and settings.
- Project switcher in dashboard sidebar
- Project settings page — rename, update description, manage stack doc
- Delete project with confirmation (cascades workspace and job records)
- Recent projects surface on dashboard home
- Per-project generation job history with status, duration, and error log
Memory isolation + concurrent job fixes
Patch release addressing two production issues found after the initial launch: memory bleed between concurrent generations and a race condition in job completion handling.
- Fixed: process memory not fully released after multiprocessing.Process exit on Windows
- Fixed: race condition where two concurrent jobs could share the same temp directory
- Fixed: SSE connection dropping silently on long-running generations (> 60 s)
- Fixed: job status stuck in "processing" after engine crash — now times out and retries
Initial launch
Logic64 ships. Governance-first AI workspace generation is now available in early access. Write your spec, generate your workspace, pull it locally.
- Web dashboard at app.logic64.com — create projects, write architecture specs, trigger generation
- AI reasoning engine (Python + FastAPI) with architecture doc parsing and constraint enforcement
- Workspace generation for Node.js + TypeScript + Fastify and Next.js 14 App Router stacks
- Authentication via Clerk — email, Google, and GitHub sign-in
- Hobby plan (free) live — 3 workspace generations per month
Stay up to date.
Get an email when something ships. No marketing, no newsletters.