CMP42 // PRODUCT // OVERVIEWTHREE MOVEMENTS · ONE GRAPH
// Product · Overview

THE SCHEMA
BUILT FOR AGENTS.

CMP42 is a headless CRM in three deliberately small parts — Context, Models, Processes — exposed through MCP tools and a fast, boring REST + GraphQL API. Built knowing the primary reader will be an agent, and the primary writer a human helped by an agent.

// The bet in brief

CRM IS A GRAPH PROBLEM NOW.

Ask a sales manager how they check the pipeline. A year ago the answer was “I open HubSpot.” Today the honest answer is “I ask my agent to summarise it, and if the summary looks off, I open HubSpot.” The workflow has moved. The database has not.

HubSpot, Salesforce and Pipedrive were designed on the assumption that a human would open the UI. Every field, every pipeline, every dashboard exists because someone was going to click on it. When an agent reads the same schema, it treats it like a chat history: it hallucinates the missing bits, invents fields that were never declared, fabricates a stage that is not in the pipeline. Not because the model is bad — because the schema was not built for the reader.

CMP42 is what happens if you build a CRM for that reader. Structured context in, structured actions out, every write attributable. We ship no proprietary AI layer: CMP42 gives your agent grounded context, and your agent produces the insight. Read the full argument →

// How it fits together

AGENTS ON TOP.
THE GRAPH UNDERNEATH.

CMP42 is layered so that every client — an MCP-native assistant, a Slack bot, the internal app your engineers ship next month — meets the same models, the same rules and the same audit trail. Read the table from top to bottom.

LayerWhat lives thereRead more
ClientsClaude Desktop, Claude Code, Cursor, Cline and custom agents built on the MCP SDK — plus Slack bots, internal apps and anything else that speaks HTTP.Integrations
SurfacesMCP tools, generated per model. REST and GraphQL with full operation parity. One auth model and one audit trail across all three.MCP tools · API
CoreContext — the typed, versioned object graph. Models — your git-versioned ontology, validated at write-time. Processes — observable workflows of agents and humans.Context · Models · Processes
OperationsThe Podma Workbench, with saved views and filters, plus a migrations CLI.Headless
RuntimeManaged by us in the EU, or self-hosted on your Postgres and object store. Same binary either way.Self-hosting

One request, end to end

  1. A partner asks their assistant to draft the renewal note for an account.
  2. The assistant calls MCP tools such as get_context and list_related to walk the graph — contacts, the deal, the last documents.
  3. It writes a draft back with write_note. The write is validated against your models; if a value does not fit, the error tells the agent how to correct itself.
  4. A process routes the draft to a human approval step before anything goes out.
  5. Every read, write, approval and token spent is recorded against the agent or human that produced it — and the run can be replayed later.
// Design Partner Program

SHAPE THE PRODUCT BEFORE GA.

Twelve companies spend six months with us in Alpha, starting October 2026, for a one-time €6,900. Direct access to the founder, a monthly co-design call, priority feature requests — and the right to propose and name one MCP tool that ships to every future customer. Applications close December 2026.

// Questions on this

What people ask here

The same answers as in the CRM FAQ, next to the case they belong to.

BAS-01

What is a CRM, and what changes when AI becomes the interface?

A CRM (customer relationship management system) keeps track of contacts, companies, sales opportunities and every interaction with them. Classic CRMs such as HubSpot, Salesforce or Pipedrive are built for people who open forms, fill in fields and drag deals across a pipeline.

Many teams now ask their AI assistant about the pipeline first and only open the CRM when the answer looks odd. When the agent reads a data model that was designed for clicks, it is quick to invent fields or stages. CMP42 turns this around: the data model is built for agents, and the interface is the AI your team already works with.

Prompts and sources for this question

HUB-06

Who is a good fit for moving from HubSpot to CMP42, and who is not yet?

A good fit Not a fit (yet)
Your team mainly reaches CRM data through agents, or wants to Your team wants a click CRM with drag-and-drop pipelines and forms
You want your ontology in Git, reviewed like code You need marketing automation with sending today
You have developers or a partner who can work with an API You need a production system today; the Alpha ships weekly releases
You need EU hosting or self-hosting You cannot run a real workload in the Alpha; then the Beta from Q1 2027 is the right time

Prompts and sources for this question

STA-01

Which teams is CMP42 meant for first?

For teams whose way of working has already moved to AI while their CRM has not. Four groups are in focus:

  • Boutique consultancies: your own framework becomes context the agent uses in the next engagement, instead of a slide copied into every deck.
  • Product-led B2B teams: agents draft renewal and follow-up notes from the real customer history.
  • Agencies with AI ambitions: you build agents for your clients that work on their real accounts instead of hallucinating.
  • Teams leaving HubSpot: you want lean data and a real API, not the next platform.

Prompts and sources for this question

Decided1
STA-05

What does the roadmap look like?

Period Milestone
October 2026 Alpha with the first design partners, weekly releases
Q4 2026 First case studies, stabilised MCP tool surface
Q1 2027 Beta with a waiting list, self-hosting image, HubSpot connector with OAuth
H1 2027 General availability (GA), published standard pricing
Q3 2027 SOC 2 Type II planned

protagx takes questions this FAQ does not answer at hello@protagx.com.

Prompts and sources for this question

Decided1

All 43 questions in the CRM FAQ

¹ Planned for a future release.