Project-owned Markdown
Keep canonical knowledge in a format you can read and version, with stable identities and immutable version history.
The product
Keep the knowledge.
Carry it forward.
A persistent memory workflow for AI-assisted development. Memora’s local-first foundation helps agents recall project context, capture lessons, and keep durable knowledge reviewable.
From a useful moment to useful memory
A deliberate knowledge loop.
- 1
Recall relevant context
Retrieve authorized project knowledge and pack it into bounded context with source references.
- 2
Capture a candidate
Record a verified lesson or observation as evidence, separate from canonical memory.
- 3
Inspect a proposal
Accuracy workflows analyze selected evidence and propose changes for governed review.
- 4
Publish reviewed knowledge
Accepted changes pass publication controls. History, diffs, and rollback support later inspection.
The engineering foundation
Readable. Portable. Inspectable.
Memora is built around Accuracy Memory, the local-first runtime in this project. Its canonical memory stays in project-owned Markdown; indexes and caches are rebuildable projections.
CLI and MCP
Use the memory runtime from a terminal or an explicitly configured MCP host. Project guidance supports Codex, Claude Code, and Cursor.
Evidence and review
Inspect proposals, preserve counter-evidence, and keep unreviewed observations separate from durable project knowledge.
For developers
Start with your project.
The runtime works from an initialized repository. Guided setup lets the owner review project guidance integrations. MCP host registration is a separate, explicit step.
These commands illustrate the workflow after installing or building the runtime and provisioning the required local authority.
Discuss setup and accessTerminal · Example commands
memory init --guided
memory --json recall \
--query "what did we decide?"
memory doctorExample commands, not a hosted live demo. Setup and available operations depend on local authorization and configuration.
Where we are
A clear view of today and tomorrow.
We are building toward shared memory for enterprise AI. The current foundation focuses on governed project memory.
| Area | Current foundation | Longer-term direction |
|---|---|---|
| Deployment | Local-first project runtime | Broader organizational memory workflows |
| Interfaces | CLI, MCP, and supported project guidance | Additional integrations as they are validated |
| Knowledge quality | Candidate capture, evidence, proposals, review, and history | Continued evaluation of memory usefulness |
| Commercial access | Contact-led discussions; no hosted checkout on this site | Availability and plans to be announced |
Your workflow, with more continuity