Skip to content
Monopea
The governed autonomous agent→

Meet Monopea,yourAI back office.

Share your workspace or work on your own with Monopea. It plans, works and keeps working inside your real documents, spreadsheets, decks, inbox and calendar — and every action that leaves the workspace waits for you.

Powered by the models you choose — routed under one governance layer

  • Qwen
  • Mistral
  • Infomaniak
  • Anthropic
  • OpenAI
  • Gemini
  • GLM
  • DeepSeek
  • Llama

How it works

From one sentence to work that is done.

You say what you want. It does the work inside the workspace, and stops at the door for anything that leaves.

Reconcile this week’s invoices|

01 · ASK

Say it once.

A job, a goal, or a schedule. It reads your mail, files and calendar to work out what the job actually is.

Proposal

Pay 3 supplier invoices · CHF 4,210

ApproveReject

02 · APPROVE

It asks at the door.

Anything above your risk threshold becomes a proposal with its exact arguments. The run waits until you decide.

  • proposed
  • approved
  • dispatched
  • logged

03 · DONE

And it is on the record.

Every proposal, decision and dispatch is written to the activity log — who, what, when, and with which arguments.

The suite

One workspace. Thirteen apps. One agent.

All thirteen are included free with every Monopea workspace — the free plan too, no card, no per-app licence. Real .docx, .xlsx and .pptx, PDFs, mail, calendar, meetings and channels, operated by one agent through 243 built-in tools. The apps are free; only model usage is paid.

DocumentsWrite, review and export real .docxExplore SpreadsheetsFormulas, charts and macros in real .xlsxExplore PDFPages, forms, merge & stamp — a real .pdfExplore PresentationsSlides, speaker notes and present modeExplore NotesA note in one of your channels — the canvas a thread is worked onExplore MailA real inbox — accounts, threads and searchExplore CalendarEvents, availability and what the agent is booked intoExplore MeetingsRecord, transcribe and summarise — on your deviceExplore ChannelsThreads, canvases and notes your agent works inExplore TeamsPeople and agents, side by sideExplore LibraryFiles, memory, the graph and the evidence behind themExplore CatalogEvery MCP tool, gated by defaultExplore Agent flowsDesign the team, the supervisor and what runs on its ownExplore 243agent tools, one gateSee the suite

Under the hood

Built to be trusted, not just to demo.

The prompt is the easy part. Underneath it, every run lands on a governance layer that decides what the agent may do on its own — and what it has to ask for.

MCP catalog

Plugs into the apps you already run

Connect any MCP-compatible server, or turn API documentation into a tool library of your own. New tools start gated, and credentials stay in the vault as {{secret:NAME}} references.

Learn more

Approval gate

Nothing leaves without you

An outward action is never executed inline. It becomes a proposal with its exact arguments, the run blocks on you, and only an approval dispatches it. Unknown tools fail closed to review.

Learn more
Proposal · send renewal quotedrafted in the workspacecredentials by {{secret}} onlyabove your threshold → askwaiting…Approveactivity logdispatched

Residency

Stored in Switzerland. Processed in the EU.

Conversations, memory, the knowledge graph, documents and encrypted secrets are stored in Zurich. Processing runs on EU infrastructure — and on the Swiss track, inference stays in Switzerland too.

Learn more
STORED · CHPROCESSED · EU

Audit trail

Every action on the record

Reads, writes into the workspace, and the outward call held for you — each one logged with who asked, what ran and what it touched, live while it happens.

Learn more
activity · your-workspaceLIVE09:12mail__readread09:19sheet__patchwritten010:26doc__insert_tablewritten010:33calendar__find_free_timeread011:40deck__add_chartwritten011:47mail draft → clientheld for you

One workspace

Your whole office, in one tab

Documents, sheets, decks, mail, calendar and meetings live in the browser, side by side with the agent that operates them. Nothing to install, on any machine.

Learn more
you.monopea.com/dashboardproposalApprove

For developers

A governance harness with an API.

Everything the dashboard does, the API does: send work, stream the reasoning, decide proposals, delegate, and manage secrets the model can never read.

  • POST /v1/chatOne full turn — retrieval, tool loop, proposals.
  • wss://…/v1/wsThe reasoning live, steerable and stoppable mid-run.
  • /mcp/:keyYour tool libraries over MCP, on scoped revocable tokens.
  • POST /v1/delegateHand work to background sub-agents and watch the tree.
  • PUT /v1/secrets/:nameA vault with no read path — not for you, not for the model.
Read the developer docs
quickstart.sh
# one turn — nothing outward runs without you
curl https://monopea-runtime.fly.dev/v1/chat \
  -H "Authorization: Bearer brain_live_..." \
  -d '{"message":"Reconcile this week's invoices"}'

# → "stopReason": "blocked_on_user"
#   the drafted calls wait in result.proposals

Pricing

Priced for founders, not enterprises.

Start free with the approval gate on. No per-action surprises — nothing is charged per proposal, per approval or per document.

Every plan, compared

Solo

For one person putting the agent to work every day.

$20/mo

6,000 tokens a month

Start free
  • One governed agent brain
  • Approval gate on mutation-capable actions
  • Write-only secrets vault
  • Long-term memory, goals & skills
  • Full audit trail & live activity

Founder

For the solo operator delegating real work.

$99/mo

29,700 tokens a month

Start free
  • Everything in Solo
  • Five times the document storage
  • Room for what real delegation piles up
  • Full MCP catalog access

Team

Most popular

For a small team operating in parallel.

$349/mo

104,700 tokens a month

Start free
  • Everything in Founder
  • Up to 5 seats with shared memory
  • Full MCP catalog access
  • Autonomy threshold per workspace & per project
  • Sub-agents in parallel — all gated
  • Priority agent runtime

The meter

The plan buys the workspace. Model usage is prepaid, in tokens.

Every application is included and nothing is charged per proposal, per approval or per document. What the models produce draws on a token balance at 300 to the dollar, topped up by each paid invoice. A generation is admitted against the balance before it runs, so at zero it stops with a clear message rather than billing you afterwards — and on the free plan the product is free while the inference is not, so start there and add a balance when you are ready.

How billing works

Scale

For companies wiring the agent into their own stack and their client work.

  • Everything in Team
  • Bring-your-own MCP tools
  • Custom skills & private memory
  • SSO & audit log export
  • Dedicated onboarding
Talk to us

FAQ

Questions people ask first.

Security in detail
Can the AI read my secrets?
No. The vault is write-only: you PUT a secret once and it can never be read back — the API returns only its name, description, and last four characters. The agent references {{secret:NAME}}; the plaintext is substituted at dispatch and scrubbed from tool results before they re-enter the model’s context.
What happens when the agent wants to do something risky?
Mutation-capable tool calls are never executed inline. Each becomes a pending proposal, the run blocks with status blocked_on_user, and you approve or reject with the exact arguments in front of you. Unknown tools fail closed to review, and every proposal, decision, and dispatch is written to the audit log.
Where is my data stored?
Your persistent data — conversations, memory, knowledge graph, documents, encrypted secrets — is stored in Zurich, Switzerland. Processing runs on EU infrastructure: the agent runtime in Paris, the dashboard and API on Cloudflare’s edge. If you choose the Swiss model track, inference also stays in Switzerland on Infomaniak-hosted models.
Which models can I use?
Claude, GPT, Gemini, Mistral, GLM, DeepSeek, Qwen, and Llama are all selectable — the governance layer is identical whichever you pick. Swiss-hosted open-weight models are the default on the Swiss track: Qwen3.5-397B, Mistral 24B, and Nemotron, hosted by Infomaniak, so inference can stay in Switzerland.
Is this compliant with the EU AI Act?
Article 14 requires human oversight to be an external control, not a prompt instruction. Monopea’s proposal gate was built as exactly that control, and the architecture maps to those obligations. We do not claim certification — formal certification has not yet been obtained — and your own compliance assessment still applies.
Can I stop a running agent?
Yes. You can steer a run mid-flight with a message, or stop it instantly — and the stop cascades to everything it delegated, since every sub-agent inherits the same gate. Work already completed is preserved and logged, so you can see exactly what happened up to the moment you stopped it.

Hand it the first job this afternoon.

Open a workspace, give the agent a task, and approve what it proposes. Stored in Switzerland, processed in the EU.

Start freeSee pricing
Monopea — your AI back office