Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.kb2b.app/llms.txt

Use this file to discover all available pages before exploring further.

kb2b lets you organize what you ingest into the POT with two complementary systems: projects (one per meeting or document) and tags (up to 19, free-form). Neither creates new POTs — everything still feeds the account’s POT. What changes is how you can filter later and what provenance each extracted fact carries.

​
Projects

A project groups meetings and documents tied to a specific initiative or phase. Examples:
  • “Acme migration” — all meetings and specs from an implementation project
  • “Q2 2026 renewal” — documents and conversations about a contract renegotiation
  • “Customer X onboarding” — discovery meetings plus the customer’s documents

​
Rules

  • 1 meeting or document = max 1 project. You can’t assign two.
  • Org-scoped. Projects live at the workspace level, not per POT. Other teammates see the same list.
  • Free-form name (up to 100 chars), auto-generated slug (lowercase, no accents, hyphenated).
  • Optional color to visually distinguish them in lists.
  • Archivable but not deleted — facts already extracted keep their provenance (project:slug) in SciPot even if the project is archived.

​
How to create one

From the picker that appears when you upload a document or import a meeting: click ”+ New project” → type a name → confirm. Created instantly and available for future assignments.

​
Tags

Tags are free-form, reusable labels with no hierarchy. Unlike projects (1:1), a single document or meeting can carry up to 19 tags.

​
Convention suggestion: prefixes

To keep tags discoverable and consistent, use category prefixes:
PrefixForExamples
TEAM/Owning teamTEAM/sales, TEAM/operations
CLIENT/Client or accountCLIENT/acme, CLIENT/distribuidora-x
CATEGORY/Content kindCATEGORY/retro, CATEGORY/discovery
TYPE/Document natureTYPE/contract, TYPE/email, TYPE/transcript
This isn’t mandatory — kb2b accepts any tag — but it keeps things sane once you have >20 tags.

​
Rules

  • Org-scoped like projects.
  • Max 19 per document or meeting (plus the project = 20 metadata sources per item).
  • Auto-generated slug from the name.
  • Optional color.

​
What happens when you ask something

When kb2b extracts facts from your documents and meetings, each fact inherits the source tags: 
source.tags = ["project:acme-migration", "tag:team-sales", "tag:client-acme"]
That means later you can filter queries by source: “Only answers based on facts from the Acme migration” or “Only what the sales team contributed”. The project: and tag: slugs are immutable — if you rename a project later, old facts still point to the original slug (this is deliberate; it preserves audit history).

​
When NOT to use this

  • If your workspace has 1 single account and few meetings, you don’t need projects. Start with tags only.
  • If your team isn’t disciplined about tagging, avoid creating hundreds of prefix-less tags — it becomes noise that filters nothing. Better few consistent tags than many chaotic ones.

​
Difference vs. workspace and POT

EntityWhat it groupsWho manages it
WorkspaceEverything: users + POTs + projects + tags. 1 workspace = 1 organization in kb2b.Workspace owner.
POTThe knowledge (facts, edges, constitution) of one account or domain. 1 account = 1 POT.The POT’s Human Curator.
ProjectMeetings and documents from an initiative. Lives at the workspace level, not per POT.Any member with edit permissions.
TagFree-form reusable labels. Workspace-level.Any member.
Page in progress Phase 1 — more screenshots and examples coming in Phase 2.