Product

Projects & ground truth

A project is where a team of people and agents works on one thing. It says who's on the team, holds the material, decides what goes in, and declares which documents are authoritative for the work, so every agent in it answers from the same brief.

A project is a team, not a folder

A project does four separate jobs that are worth keeping apart, because they compose in one direction only:

The sentence the whole model rests on: a project-visible artifact is readable by the members of any project whose collection holds it. Membership and collection, both, and nothing else.

What a member actually gets

A member can read the project, and the project-visible artifacts in its collection. That is the entire grant. Stated negatively, because people reliably assume more: no writes — not the settings, not the brief, not the collection — and no memory access at all. Memory is personal; putting an agent on your project gives it none of your memories.

A delegate adds four verbs — add or remove members, attach a sub-project, rename the project's display label, and approve or reject a member's proposed addition — and is still not a writer: it cannot add to the collection directly, declare ground truth, or change a retrieval setting. Only the owner may grant it, so the role cannot propagate itself through a team. Anyone may remove themselves, unconditionally. Since v1.18.1 a delegate of a parent project can also read its sub-projects as the owner does, without being able to change them.

Members propose, owners decide

A member can't change what a project holds, but it can ask. Proposing an artifact puts it into the project's pending additions; the owner or a delegate opens it, then approves or rejects. An approved artifact joins the collection and records who added it. A rejected one is left untouched, so it can be fixed and proposed again, and the proposer can withdraw a proposal at any time.

A pending proposal grants nothing. Until someone approves it, no one but the owner and the delegates who decide it can read it, and only there: it isn't in recall, and no other member learns it exists. A proposal is a request, and a request that leaked its contents would be a publication.

Agents use the same flow, over MCP: an agent proposes, and an agent with the authority to decide can list what's pending and approve or reject it. Members can also flag an artifact the collection already holds, which proposes its removal.

Units of work

A repository is rarely one body of work. A sub-project is a unit inside one — a service, a component, a stream someone owns — and the important part is that a project is the unit of ownership, a repository is not. A second person on a shared repository owns their unit rather than needing a seat inside someone else's.

A unit can draw on its parent's brief, because what is true of a repository is true of its parts. Inheritance flows down only: a unit deciding something for itself is not the repository deciding it. And prefer declaring in each unit over declaring high and opting out — a claim on the parent that is only true for two of five units is a claim about the repository that is false.

A room to talk in

A project can have one or more chat channels bound to it, so the team has somewhere to discuss the work. A binding is addressing, not access: it grants nobody anything in ENGRAM. What it adds is a doorbell. When a member proposes something, the owner and delegates are rung in the channel; when it's decided, the proposer is rung back. A sub-project shows its parent's channels unless its owner turns that off. Channels & doorbells →

Retrieval finds what's relevant. It doesn't know what's true.

Associative recall is good at reaching for what relates to a question. It has no opinion about which of two related documents is the one your team actually decided on. Point five agents at a knowledge base containing last quarter's design, this quarter's revision, and a colleague's exploratory note, and they will each answer from a different one — confidently, and without telling you which.

Ground truth is the missing claim. A project declares which articles are authoritative for it, and retrieval performed under that project prefers them. The claim being made is “this is true” — not “this is relevant”, and not “we are working on this”. That distinction drives everything else on this page.

Declaring a brief

Declaring replaces rather than appends: each call defines the complete set, so re-declaring the same articles changes nothing and declaring an empty set clears the brief. It is idempotent by construction, which makes it safe to put in a setup script that runs more than once.

The guard on what may be declared is readability: can this project's members read the article? A brief they cannot read is not strict, it is broken — a success response and no effect, which you discover months later. So a private article is refused loudly rather than accepted quietly.

A team's brief no longer has to be published to the world to be shared with the team. An article set to project visibility, in a project whose collection holds it, is readable by that project's members — and that is enough to declare it. Before, public was the only visibility that satisfied the rule, so anchoring a team meant publishing to everyone.

A declared article stays authoritative across edits: revise it and the newest version is what answers, with no need to re-declare.

Prefer, or bind

Two policies, and the difference is how much room the agent has:

There is also a negative list: articles that enumerate false claims. Note the shape carefully — a refuted article is one that lists things which are wrong, not an article that is itself wrong. It is a deliberately separate kind of link rather than a flag on the same one, so that a single missing filter can never cause the list of known falsehoods to be read as ground truth.

Workers inherit it automatically

This is the most commonly misunderstood point, because people go looking for a “use ground truth” flag on the call and there isn't one. An agent declares which project it is working in — by id, or by the repository URL, which ENGRAM resolves — and the brief follows. Nothing else is passed, and nothing has to be threaded through a fan-out.

Two things that sound alike and are not: membership grants the read — a member can read a project-visible artifact whether or not they declare the project — while declaring the project grants the authority, which is what makes the brief outrank ambient content. So “I got the passages either way” is not evidence the anchor does nothing; it changes what wins.

The same set checks new writing

Ground truth is not only a retrieval preference. When ENGRAM drafts an article to close a gap it found in what it holds, it verifies the draft's claims about how ENGRAM works against a maintained ground-truth corpus, and flags or rejects the ones that contradict it. Checking a draft against a project's own declared brief is the next step. A memory layer that lets agents write into it needs an answer to “what stops them writing something untrue?” — this is where that answer starts.

Private beta — by invitation

If you're running agents that work together and want a memory and a project space they can share on purpose, tell us what you're building.

Get in touch → info@pvelua.net