Questions
Everything people ask, in one place.
What Mneia is, how it differs from a memory tool, what it costs, what happens to your data, and what is actually built today.
The product
What Mneia is, and what it is not.
- What is Mneia?
- Mneia is the shared project memory and handoff layer for teams working with AI agents. It does three things: checkpoint, which captures the decisions, constraints, and open questions out of a session at a task or day boundary; rehydrate, which assembles the minimal high-signal context slice for the next task under a token budget; and handoff, which produces a receivable artifact when work changes hands.
- How is Mneia different from an AI memory tool?
- Memory tools give you a place to store context and a way to query it, which leaves the burden on whoever is picking the work up. Mneia produces a handoff: an object a person or an agent receives, stating what is done, current state, open questions, constraints, and the single next action. It also carries a superseded section recording what was tried and rejected, so a fresh agent does not propose the approach the team already ruled out.
- Why is context compaction a problem?
- Compaction is lossy by design and task-blind. It fires when the context window is full, which is the worst possible moment, and it produces nothing you can review. An agent that has been compacted can silently lose a constraint established two hours earlier and confidently propose an approach that was already rejected.
- What happens when a human and an agent disagree?
- The human wins, always. An agent assertion that contradicts a human-confirmed item is stored as disputed rather than silently applied. Two humans disagreeing is never auto-resolved, because arbitrating between two people is not a decision software should make on their behalf.
- Which AI coding tools does Mneia work with?
- Mneia ships an MCP server that works in Claude Code, Cursor, Codex, or any MCP client, plus a CLI and file interop with the AGENTS.md and CLAUDE.md files you already keep. The inner loop stays in your terminal, because a handoff that only works inside one vendor tool is a session feature rather than a handoff.
- Does Mneia replace my agent framework or observability stack?
- No. Mneia sits beside the frameworks and observability tools you already run, never above them. It is not an agent orchestrator, not a tracing or evals product, not an enterprise document search, and not an agent of its own. Everything it builds serves checkpoint, rehydrate, and handoff.
Getting started
Access, install, and what changes about your day.
- Is Mneia available yet?
- Not generally. Mneia is in staged early access and the waitlist is the way in. The client packages are not published to npm yet, so the install steps in the documentation describe what happens once your account is enabled rather than something you can run today. Everything the documentation describes about the command surface and the tool surface is real and built.
- How do I get access?
- Request it from the form on the home page. We are onboarding teams in stages and will email you when your account is enabled. That email is the only one the waitlist sends you.
- What do I install?
- Two packages, and most people want both. The CLI, @mneia/cli, gives you mneia init, brief, checkpoint, log, and status in the terminal. The MCP server, @mneia/mcp-server, exposes mneia_rehydrate, mneia_assert, mneia_checkpoint, and mneia_search to Claude Code, Cursor, Codex, or any other MCP client, so the agent can read and write project memory without you relaying it.
- What does mneia init actually do to my repository?
- Three things. It writes .mneia/config.json binding the directory to a workspace and a project. It reads the AGENTS.md, CLAUDE.md, and .cursor/rules files you already keep and imports the constraints it finds in them, so the project does not start empty. And it writes a generated section into AGENTS.md, inside a fence it owns, so the agent sees the binding without you configuring anything twice. Nothing outside that fence is touched.
- Do I have to change how I work?
- The inner loop stays in your terminal and your editor. In practice you add two habits: call rehydrate at the start of a task, and checkpoint at the end of one. Both can be done by the agent through the MCP server rather than by you, and the CLI exists for the moments you would rather do it yourself.
- Does Mneia read my session transcript automatically?
- No. mneia_checkpoint records a batch of candidate items that the agent extracted; it does not read the transcript itself. That is deliberate. Ambient capture produces noise, and capture at an explicit boundary produces something a human can review before it becomes project memory.
Pricing and billing
What is free, what is metered, and what is not.
- How much does Mneia cost?
- Solo is free. Team is $24 per user per month and includes a checkpoint allowance sized well above ordinary use. Enterprise is custom priced, for organisations that need governance over what context agents can see. Pricing is in preview: the seat price is not final until we have measured what a real checkpoint costs to run, and if it moves it moves before anyone is billed.
- Is the free tier a trial?
- No. Solo is free and stays free. Individual use is how the product spreads, so charging for it would be charging for our own distribution. It covers one project, 30-day history, capped checkpoints, the MCP server and CLI, and handoffs to yourself.
- What exactly is metered?
- Only the checkpoint, because the extraction call it runs is effectively the entire marginal cost. Contradiction detection is rolled into the checkpoint. Rehydrate is one indexed query costing fractions of a cent and is not metered. Handoff, log, status, and search are not metered. Storage carries a fair-use ceiling only.
- Do I need to bring my own model provider API key?
- No. We pay for inference, not you. Charging a seat price and then asking you to fund the model calls on top would be charging for the same product twice. The seat price carries the variable cost, which is why the included checkpoint allowance is a real number rather than a formality.
- What does the Team plan add over Solo?
- Shared projects and roles, cross-team scope, conflict resolution between humans and agents, unlimited history, team handoffs, and the web review app.
Data, privacy, and licensing
Where your content lives and who can reach it.
- Do you train models on my content?
- No. Your project content is not used to train models, ours or anyone else’s. The Privacy Policy states this as an obligation rather than an intention.
- Can I self-host Mneia?
- No. Mneia runs as a hosted service, and the CLI and MCP server require an account and do not function without it. We would rather say that plainly here than have you discover it after installing. Bring-your-own-cloud deployment is an Enterprise conversation, not a download.
- Is Mneia open source?
- Partly, and the line is drawn on purpose. The client packages are Apache 2.0: the CLI, the MCP server, and the core, which holds the schema definitions, the handoff format, the extraction prompts, and the ranking algorithm. The hosted API, the store, billing, and the review surfaces are proprietary.
- Where is my data stored?
- In managed Postgres in the United States. Every row carries the workspace it belongs to, and the database enforces workspace isolation with Postgres row-level security rather than relying on the application to remember. The Privacy Policy lists every subprocessor that touches your data and what each one does.
- What telemetry does Mneia collect, and can I turn it off?
- Mneia records structured product events — that a checkpoint ran, how large a rehydration slice was, whether an item was confirmed — because the quality of the extraction cannot be improved without measuring it. Your content is not in those events by default, and you can opt out entirely by setting MNEIA_TELEMETRY=off in the environment of the CLI or the MCP server.
- What happens to my project memory if I stop paying or leave?
- It is yours. The record of what your team decided and why is the thing that accumulates value, so it is exportable rather than hostage. The Terms set out the retention window and the export path.
Support
How to reach a person, and how quickly.
- How do I report a security problem?
- Email security@mneia.dev with enough detail to reproduce it. Please do not include anyone else’s data in the report. Mneia is small and pre-general-availability, so that address reaches a person directly rather than a queue.
- How do I exercise a privacy right — access, correction, or deletion?
- Email privacy@mneia.dev. Requests under the GDPR, the CCPA and other US state laws, and India’s DPDP Act are answered within the statutory window set out in the Privacy Policy. If you are on the waitlist and want off it, the unsubscribe link hard-deletes your address rather than flagging it.
- Is there a support SLA?
- Not before general availability, and we would rather say so than publish a number we cannot hold. Enterprise plans carry a support SLA as a contracted term. Everything else gets a reply from a person as soon as we can manage it.
Not answered here
The documentation is more specific.
Help covers the errors people actually hit. The reference pages carry every flag and every tool. And if something here is wrong, that is a bug worth telling us about.