Skip to content

docs: document organization prompt cards, team scoping, and token usage - #3109

Merged
benjaminshafii merged 1 commit into
devfrom
docs/team-prompts-and-token-cost
Jul 25, 2026
Merged

docs: document organization prompt cards, team scoping, and token usage#3109
benjaminshafii merged 1 commit into
devfrom
docs/team-prompts-and-token-cost

Conversation

@benjaminshafii

Copy link
Copy Markdown
Member

Why

Two documentation gaps from enterprise rollout questions, both cases where the product ships a behavior the docs never mention.

1. Organization prompt cards + per-team scoping

cloud/share-with-your-team/desktop-policies.mdx documented only the boolean toggles and never mentioned the custom prompt feature at all, though it has shipped. New cloud/share-with-your-team/team-prompt-cards.mdx (23 lines) documents, all verified against code:

  • Limits: 2 or 3 prompts, prompt max 500 chars, description max 120 chars.
  • The counter-intuitive mapping: the admin's Description becomes the card title; the Prompt text becomes the card body and the composer text. Clicking fills the composer, it does not auto-send.
  • Per-team / per-person scoping via memberIds/teamIds — the concrete answer to "different prompts for different departments".
  • The footgun: boolean policies are OR-unioned, but prompts are winner-takes-all by priority, then createdAt, then id.
  • The default policy cannot be assigned, so a department-specific set needs a non-default policy.
  • Requires Enterprise (desktopPolicies entitlement, HTTP 402), and changes arrive via GET /v1/me/desktop-config so a member may need to reload.

2. What actually uses tokens

No page anywhere in packages/docs covered cost, tokens, or usage — and enterprise buyers sizing thousands of seats ask this first. New start-here/do-work-with-it/what-uses-tokens.mdx (23 lines) states only what is true today: every run costs, there is no result cache; cost scales with context read and tool calls, not with how many prompt cards an admin configures; configuring/installing/publishing costs nothing until someone runs something; a skill can reduce cost by making a task deterministic.

Deliberately omitted because they do not exist in this repo: compiled widgets that execute without the model, result caching, free dashboard refresh, and scheduled runs. Documenting any of those would be a promise we cannot honor.

Verification

docs.json parses, both pages registered and resolve on disk, all internal links resolve, no image paths invented. Code citations for every numeric claim are in the commit discussion.

Docs-only, no runtime path — no fraimz.

@vercel

vercel Bot commented Jul 25, 2026

Copy link
Copy Markdown
Contributor

The latest updates on your projects. Learn more about Vercel for GitHub.

Project Deployment Actions Updated (UTC)
openwork-app Ready Ready Preview, Comment Jul 25, 2026 1:00pm
openwork-den Ready Ready Preview, Comment Jul 25, 2026 1:00pm
openwork-den-worker-proxy Ready Ready Preview, Comment Jul 25, 2026 1:00pm
openwork-diagnostics Ready Ready Preview, Comment Jul 25, 2026 1:00pm
openwork-landing Ready Ready Preview, Comment, Open in v0 Jul 25, 2026 1:00pm

@mintlify

mintlify Bot commented Jul 25, 2026

Copy link
Copy Markdown
Contributor

Preview deployment for your docs. Learn more about Mintlify Previews.

Project Status Preview Updated (UTC)
differentai 🟢 Ready View Preview Jul 25, 2026, 1:01 PM

💡 Tip: Enable Workflows to automatically generate PRs for you.

@mintlify

mintlify Bot commented Jul 25, 2026

Copy link
Copy Markdown
Contributor

Preview deployment for your docs. Learn more about Mintlify Previews.

Project Status Preview Updated (UTC)
differentai 🟡 Building Jul 25, 2026, 1:00 PM

💡 Tip: Enable Workflows to automatically generate PRs for you.

@benjaminshafii
benjaminshafii merged commit e61c527 into dev Jul 25, 2026
12 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant