Connected sources

GitHubSynced

SlackSynced

Google DriveSynced

NotionSynced

ConfluenceSynced

PostgresSynced

Compilation pipeline

slack-threadrunbook.mdpricing-notesrfc-114

extract · deduplicate · structure

context compiler

write

context/company/pricing.mdnew

context/ — Open Knowledge Format

overview.md

pricing.md

onboarding.md

macros.md

## Pricing policy

Enterprise discounts require approval below an annual contract value of $25k…

Source: pricing-notes, rfc-114 · Updated 2 days ago

Pull request

Merge duplicate onboarding notes

maintenance-agent wants to merge 1 file into main

- Last updated March 2024. Prices on request.

+ Reviewed 2026-08. See pricing.md for current tiers.

Checks passed · awaiting your review

mcp.json

{
  "mcpServers": {
    "querypanel-context": {
      "url": "https://mcp.querypanel.com/org_123"
    }
  }
}

ClaudeCursorChatGPTCodex

One file, every team

context/company/*

Engineering

How do we rotate API keys?

Support

What is the refund window?

Sales

When does SOC 2 expire?

Security

Who owns incident response?

Same reviewed source behind every answer.

01 / 06The problem

Your company’s knowledge is scattered.

Architecture decisions live in Slack threads, onboarding notes in Google Docs, runbooks in a repo nobody updates. Every agent inherits the gaps.

GitHub
Slack
Drive
Notion

Join the Waitlist Explore the system

02 / 06Context compilation

QueryPanel compiles it into knowledge files.

Connect the systems you already use. QueryPanel continuously extracts what matters, resolves duplicates, and writes structured Markdown your team can review.

Extract
Deduplicate
Structure

03 / 06Open Knowledge Format

Plain Markdown, versioned in Git.

The compiled output is files: readable in any editor, reviewable through pull requests, and portable out of QueryPanel at any time.

Markdown
Diffs
Pull requests

04 / 06Kept current

Agents maintain the knowledge base.

Background agents flag stale pages, merge duplicate notes, and open pull requests with proposed fixes. You approve every change; nothing rots quietly.

Deduplicate
Refresh
Repair

05 / 06MCP delivery

Every agent reads the same context.

Expose the compiled knowledge once over MCP. Claude, Cursor, ChatGPT, and your own agents all query the same reviewed source.

MCP
Open delivery
Custom agents

06 / 06One source of truth

Stop re-explaining your company to software.

One maintained context layer behind every agent, instead of five retrieval pipelines drifting apart.

Agent-first
Company-wide
Continuously maintained

Join the Waitlist

Scroll to compile

Why now

Every team is deploying agents. Almost none share context.

The models are capable. What they lack is a shared, current understanding of your business.

Context lives everywhere.

Knowledge changes every day.

Each agent rebuilds the same retrieval.

Teams need one trusted source of truth.

A context compiler

Knowledge infrastructure, not another search box.

QueryPanel turns fragmented source material into an open, reviewable layer designed for both people and agents.

Open Knowledge Format

Company knowledge becomes structured Markdown files: readable by people, editable by agents, versioned in Git.

Human-readableAgent-editableReviewable diffs

Continuous compilation

Connect the systems your team already uses. QueryPanel extracts, organizes, and refreshes knowledge as it changes.

GitHub and docsDrive and PDFsSlack and Notion

MCP ready

Expose one trusted context layer to the agent tools your teams are already adopting.

ClaudeChatGPTCursor and Windsurf

AI knowledge maintenance

Background agents do the upkeep that makes company knowledge dependable: merging duplicates, flagging stale pages, opening reviewable pull requests.

Merge duplicatesDetect stale docsPropose fixes

How it works

From scattered inputs to dependable context.

01
Connect
Connect the repositories, workspaces, documents, and systems where company knowledge already lives.
02
Structure
QueryPanel extracts useful knowledge, resolves overlap, and organizes it around how your company actually works.
03
Compile
The result is written into an open, filesystem-based knowledge format with history and reviewable changes.
04
Serve
Your agents consume reliable, current context through MCP instead of rebuilding retrieval for every tool.

A different philosophy

Search stores what you wrote. A compiler maintains what agents need.

Traditional knowledge bases

Store documents
Depend on manual updates
Optimize for search
Serve human readers first

QueryPanel

Compiles knowledge
Uses AI-assisted maintenance
Designed for agents
Writes an open format
Ships context through MCP

One shared understanding

One context layer behind every agent.

Engineering, support, security, and sales all read from the same maintained source instead of rebuilding retrieval per tool.

EngineeringSupportSecuritySalesInternal tools

Early access

Help shape the context layer your agents need.

We are inviting engineering leaders, AI platform teams, and technical founders who are already building agent workflows.

Discuss your current context architecture

Influence connectors and deployment priorities

Get invited to private product reviews

FAQ

The practical questions.

What is Open Knowledge Format?

A portable, filesystem-based way to represent company knowledge as structured Markdown plus metadata and links. People can inspect it, Git can version it, and agents can use or improve it without a proprietary reader.

Why not just use RAG?

RAG finds passages at query time. It does not reconcile duplicates, preserve a reviewable history, or keep shared context coherent across many agents. QueryPanel compiles and maintains the knowledge layer that retrieval systems consume.

Can I self-host it?

The knowledge output stays open and file-based. We are working with early design partners to shape deployment and self-hosting requirements rather than promising one operating model too early.

Does it replace Notion?

No. Notion remains where people write and collaborate. QueryPanel connects to systems like Notion and compiles their useful knowledge into a consistent context layer for agents.

How is this different from Glean?

Glean optimizes enterprise search for people. QueryPanel compiles an open, filesystem-native knowledge layer that agents consume and maintain, with portability as a core requirement.

Does it work with Claude and ChatGPT?

Yes. Delivery is built on MCP, so the same compiled knowledge reaches MCP-compatible clients such as Claude, ChatGPT, Cursor, Windsurf, Codex, and your own agents.