Prism Cursor for product managers · open source An AI workspace that proposes, never overwrites

One beam in.
The spectrum out.

Product signal arrives undifferentiated — interviews, support exports, ticket dumps, metrics. Prism separates it into the few things worth acting on, and writes them into your PRDs as diffs you accept or reject.

Drag to aim
8
interconnected sources in the seeded workspace
40
raw support tickets, pivoted on request
9
agent tools, every mutation reviewable
0
changes applied without your approval
01
Walkthrough

Six steps, start to shipped.

A pass through a real workspace: Meridian, a B2B analytics company with a churn problem. A scripted recording of the actual interface — with one exception. Step four is yours to drive.

Prism Meridian TEAM Search documents, sources…⌘P ⚡ Opus 5
PRD — Saved Views×
PRD — Saved Views & Shared Dashboards PRD In review @ 💬 ✦ Edit with AI
551 words82 lines~878 tokens⌘⇧E to edit
0 pending ⚠ 1 need attention 6 docs · 8 sources Opus 5 (demo) Aayush Sharma · owner
← → step · space play

The workspace

Documents you can edit, and read-only evidence you can cite.

02
How it works

Evidence in, reviewed change out.

Four stages, and a gate at the end of every one. The order matters — this is the path a decision actually takes, and nothing skips ahead.

01

Bring the evidence in

Drag a support export, paste interview notes, connect a tracker. CSVs parse into real tables. Nothing is summarised on the way in — the agent reads the source, not a digest of it.

sources the agent can search and quote
02

Ask a question worth asking

It searches, reads, and pivots the raw data, then answers with citations. Where two framings disagree — volume versus revenue — it shows both and says which it trusts.

an answer with the excerpt behind every claim
03

Turn the answer into a change

Tell it to update the spec. The edit arrives in your editor as a diff, hunk by hunk. Take two, leave one — the merged result is what lands, and the audit trail records that it was partial.

a proposal, never a silent write
04

Ship it, or find out why you can't

Break the spec into tickets with acceptance criteria and push. If the tracker's credential is expiring, the push is blocked with the reason instead of half-writing your backlog.

tickets in Jira or Linear, or a refusal you can act on
03
The premise

Cursor is a developer tool wearing a PM costume.

Every complaint PMs have about using Cursor traces back to one fact: it was built to edit code. No team visibility. No memory between sessions. Integrations that fail silently. And a floor of codebase literacy before any of it pays off.

Prism keeps the interaction model — the tree, the inline edit, the reviewable diff — and swaps the substrate. Your repository is your product knowledge.

substrate.diff+1 −1
a repository of source files, edited by engineers
+a workspace of PRDs, interviews, tickets and metrics, edited by whoever owns the decision
In CursorIn Prism
File tree over a repoWorkspace tree over documents and evidence
⌘K inline code edit⌘K inline edit on any PRD paragraph
Agent proposes code diffsAgent proposes document diffs, accepted hunk by hunk
.cursorrulesProduct Memory — persistent, agent-writable, still reviewed
MCP servers that drop silentlyIntegrations with stored health that refuse loudly
TerminalTicket composer, pushed to Jira or Linear
04
The one rule Every change the AI makes is a proposal.

Agent edits, ⌘K rewrites, and updates to Product Memory all travel the same path: a diff in your editor, accepted or rejected hunk by hunk. One mental model for AI changes, not three — and no code path that writes to a document without either your explicit save or your approval.

That is a compliance property, not a preference. No product org will let an agent silently rewrite the spec engineering builds from.

Per-hunk

Take the good half

Accept two hunks, reject the third. The merged result is recomputed and recorded, so the audit trail shows what actually landed — not what was offered.

Snapshotted

Undo any AI change

Every accepted diff snapshots the document first. Restore from history, and the restore itself is snapshotted. Nothing is destroyed.

Cited

Claims carry receipts

Assertions link to the source excerpt behind them. When the evidence is thin, it says so instead of sounding confident.

Loud

Failures you can see

Integration health is stored and checked. A degraded connection blocks the push with a reason rather than reporting a success that never happened.

05
What's in the box

An editor's rigour, pointed at product decisions.

Ingest

Drop a CSV, get a pivot

Drag a support export onto the sidebar. It parses into a real table, and the agent pivots it into the severity-by-frequency matrix you'd otherwise build by hand.

Write

⌘K on any paragraph

Select, instruct, watch the rewrite stream in as a word-level diff. Apply or discard.

Remember

Product Memory

Strategy, guardrails and vocabulary injected into every request. The agent can add to it — as a diff you approve.

Ship

Tickets with criteria

Break a PRD into an epic and stories with Given/When/Then acceptance criteria. Edit them, then push.

Share

Links that expire

Publish a read-only page for a stakeholder. No account needed, revocable instantly, and no route back into the workspace.

Review

Comments and activity

Feedback anchored to the quoted passage, plus a workspace feed built on the audit log — so visibility and governance can't drift apart.

06
Stated plainly

What this does not do.

A page that only lists strengths tells you nothing. These are the limits that would matter to you, in the order they'd bite.

AUTH

No login. Single-tenant demo state. The session seam is isolated to one function, so wiring a real identity provider touches nothing else — but it isn't wired.

PRESENCE

No live cursors. Collaboration is asynchronous: comments, attribution, activity. Real-time presence is separate infrastructure and isn't built.

CONCURRENCY

Last write wins. Two people in one document will overwrite each other. Version history means nothing is lost, but there is no merge.

TRUST

Uploaded sources are untrusted input reaching a tool-calling model. The review gate contains the blast radius, but prompt injection via a malicious export is a real consideration before ingesting production data.

VENDORS

Integration adapters are stubs. Health checks and the push flow work end to end; no HTTP request leaves the machine yet.

07
Run it

It's open source. Clone it and go.

Prism keeps its workspace in a local database, so this page is a walkthrough rather than a hosted instance — which also means there's no waitlist and nothing to request. The seeded Meridian workspace comes up with six documents and eight interconnected sources to interrogate.

terminal
# no account, no key, no configuration
git clone https://github.com/aayusharmaaa/prism.git
cd prism && npm install
npm run dev            # → http://localhost:3000

# optional — real model output instead of the scripted agent
echo "ANTHROPIC_API_KEY=sk-ant-..." > .env.local

Without a key it runs a scripted agent that still executes real tool calls — searches, edits and ticket batches land in the review queue exactly as the live model's would. Only the prose is canned.

aayusharmaaa/prism Issues and forks welcome