Devplan raises $2.5M to turn customer feedback into shipped features. Read more on GeekWire.

ProductProduct brain

Know why everything was built.

The AI-native way to understand your product and your customers: every request, decision and line of code connected, answers with sources, and the same context for your AI agents.

Start free Book a demo

Slack and Teams

Ask in Slack. Get the answer and the source.

Anyone can ask about the product where they already work.

  • No engineer interrupted
  • Sources on every answer
  • Respects permissions
# product-questionsSlack
PR

Priya Support

@devplan how does refund handling work now?

Devplan APP

Refunds go back to the original card. Partial refunds need admin approval since Aug 3.

PR #611Decision · Aug 3Help doc

✅ 2 · 1 reply

MK

Maya Sales

@devplan can admins export audit logs?

Devplan APP

Yes. Settings → Security → Export, as CSV. Shipped in July.

PR #540ENG-377Changelog · Jul

✅ 3

One shared product brain

What changes for your team.

One live picture of your product means fewer questions, fewer meetings and fewer wrong builds.

Product knowledge graph

New hires answer their own questions

Code, calls, tickets and decisions in one place they can ask.

90% of a new hire's questions answered without a founder at Tato (Tato's estimate)

Daily briefing Radar

Start every day knowing what matters

Risks, requests and wins for your role, without digging through Slack.

See the daily digest →

Custom dashboards

Numbers without waiting on an analyst

Ask for a dashboard; it arrives in Slack or email on a schedule.

Living docs Product map

Docs that are never out of date

Written from code and conversations, updated as you ship.

Centralized decisions

Stop relitigating settled questions

Every decision captured, routed for approval and easy to find.

Goals and context

Every spec follows your strategy

Goals and company context shape every recommendation, spec and agent action.

Under the hood

Why Weaver's answers hold up.

Connecting Claude to your tools gets you a few search results. Weaver joins what your product does with what your customers need, cleans it up and keeps the source of every fact, so nothing gets missed.

  1. 1Entity normalisationOne customer, one feature, one name. “Acme”, “ACME Corp” and acme.com become one account; “SSO”, “SAML login” and “single sign-on” become one request.
  2. 2Signals add up to insightsForty mentions across calls, tickets and Slack become one ranked insight, with its demand, revenue and trend.
  3. 3Evidence for every answerLossless: everything is compressed into the graph without dropping the source, so every answer links to the exact call, ticket or commit.

Up to 6x faster

1m 42s vs 9m 57s to answer the same question: Claude with Devplan's MCP, versus Claude connected to each tool directly

Up to 4.5x lower model cost

$1.16 vs $5.38 for that same answer, because Weaver hands Claude the evidence instead of making it search every tool

From one side-by-side test in our demo video. Results vary with the question and your data.

For every team

One product brain, every team, no per-seat charge.

Product

Know what to build next, with the evidence to back it.

Engineering

Clear specs grounded in the code, and AI agents that know the context.

Leadership

Roadmap progress, risks and decisions across every team, with no status meetings.

Sales

Answer “does it do X?” and “when is Y shipping?” in Slack, without pulling in an engineer.

Customer success

Track the features your accounts asked for and tell them the moment they ship.

Support

Resolve tickets faster with always-current docs on how every feature actually works.

Marketing

Turn what shipped into launch content and release notes, without chasing the team for details.

Build or buy

Claude plus a few MCP servers, or Devplan?

Both give Claude your tools. Only one gives your whole team the same, current picture.

Claude + MCP yourselfDevplan
Remembers decisions and whyStarts from scratch each sessionKeeps the history, with sources
Connects code, customers and ticketsOne tool at a time; you join the dotsAlready linked for you
Same answer for everyoneDepends on who set it upOne shared picture for the team
Respects who can see whatYou build and audit itBuilt in, from your tools' permissions
Stays current as tools changeYour engineers maintain itManaged for you
Goes beyond answersAnswers onlySpecs, tickets, status and risks too
Compare the cost →

Use it where you already work

Connect once. Ask from Claude, Codex or Slack.

Devplan's MCP server gives your AI tools the same product brain your team uses.

  1. 1
    Connect your tools

    GitHub, Jira or Linear, Slack, Zendesk and more, in minutes.

  2. 2
    Add Devplan's MCP server

    One connection for Claude, Claude Code, Codex and Cursor. Setup guide

  3. 3
    Ask anywhere

    In your editor, in Claude or ChatGPT, or mention Devplan in Slack.

Claude CodeCursorCodex
~/app · claude-code · devplan mcp

Customer stories

Teams that ship what customers ask for.

See all stories

Questions

Product brain, answered

What is a product brain?

Devplan's product brain, called Weaver, sits at the intersection of product understanding and customer understanding: one live, shared picture of your code, tickets, docs and decisions, connected to the calls, tickets and chats where customers say what they need, and kept current automatically. Devplan builds it from the tools you connect. People ask it questions in Slack, Teams or the web app, and AI agents read it through MCP, so everyone works from the same facts.

What is a context graph?

Devplan builds a context graph: a live map that links your code, customer requests, tickets and decisions, and records why each thing happened. Your team and AI coding agents can ask it questions and get answers with their sources.

Is a product brain the same as a company brain?

Devplan's product brain is a company brain focused on your product. It connects code, customers, tickets and decisions rather than every document in the company, so it can explain why something was built and give AI coding agents the right context.

How do I give Claude Code context about our product decisions and roadmap?

Connect Devplan to Claude Code through MCP. Devplan builds a live product knowledge graph from your code, tickets, calls, Slack and decisions, and Claude Code reads it when it plans or writes code. Add a short instruction to CLAUDE.md or AGENTS.md and the agent checks Devplan automatically. Cursor and Codex work the same way.

How do I build a company knowledge graph for AI agents?

Devplan builds it for you. Connect GitHub or Bitbucket, Jira or Linear, Slack, Notion, Confluence, Google Drive, Zoom and Granola, and Devplan links features, customers, tickets, decisions and code into a knowledge graph that stays current as work happens. AI agents read it through MCP; people ask it in Slack or the web app.

Our product knowledge is spread across Notion, Slack, Linear and Google Docs. How do we make it usable by AI assistants?

Devplan connects to Notion, Slack, Linear and Google Drive, among others, and turns what's in them into one product knowledge graph with sources. Your team asks it questions in Slack, and AI assistants like Claude Code, Cursor and Codex read the same context through one MCP connection.

How can AI help new engineers understand our product faster?

Devplan answers the questions new hires would otherwise ask a founder or senior engineer: why something was built, how a feature works and what was decided. Answers cite the commit, call or decision behind them. At Tato, 90% of a new hire's product questions get a useful answer from Devplan without involving a founder (Tato's estimate).

How is Devplan different from an AI knowledge base or enterprise search?

Devplan connects what your company knows about the product, not every document in the company. Search finds files; Devplan's product brain links code, customer requests, tickets and decisions and keeps them current, so it can answer why something was built and what's at risk, and power specs and AI agents too.

Is my data secure?

Devplan encrypts data in transit and at rest, respects the permissions from your source tools, and never uses your data to train Devplan or third-party models. Enterprise teams can self-host on GCP or AWS. See Devplan for enterprise.

Give everyone the same answer.

Connect your tools in minutes. $100 in free credits, no credit card.

Start free Book a demo