How is Relm's pricing different from Attio's?

Attio, like most team CRMs, is priced per user seat. Relm is priced per API request because the caller is an agent, not a seat: Free is 1,000 requests a month with no card, Pro is $29/mo for 100,000 and Scale is $249/mo for 2,000,000, plus a free unlimited test mode. A headless, agent-driven CRM with no human seats maps naturally onto a request meter.

Relm vs Attio - Human-Team CRM vs Agent-First CRM

Q: Is Relm an Attio alternative?

They aim at different operators. Attio is a polished CRM for a sales team of people, with AI assists layered on and a UI as the primary surface. Relm is agent-first: the primary operator is an AI agent, the human dashboard is deliberately thin, and everything is designed to be discovered and driven over an API and a native MCP server. If a team of people runs your pipeline with some AI help, Attio fits; if an agent runs the CRM largely unattended, Relm is built for that.

Q: Does Attio have a native MCP server?

Attio has a capable REST API and has added AI features, and there are community MCP wrappers. Relm ships a first-party native MCP server at https://api.relmcrm.com/mcp with 37 typed tools plus relm_describe_schema, and every operation is also an A2A endpoint and an OpenAPI-described REST call. Agent access is the product in Relm, not an add-on.

Q: Can Relm and Attio be used together?

Yes. Some teams keep a human-facing CRM for the sales team and run Relm as the system their agents own - lead triage, enrichment, meeting-notes-to-CRM, follow-up sequences - syncing the records that matter between them via each product's API and Relm's webhooks. The choice is not always either/or.

Q: Why choose an agent-first CRM at all?

Because a CRM designed for people and then opened to agents still assumes a human in the loop: it fails in ways a person would catch, documents itself for developers, and prices by seat. An agent-first CRM assumes the opposite - the agent reads the whole schema in one call, every error is machine-readable with valid_options so the agent recovers on its own, and there is no seat to buy for a bot. When most reads and writes are automated, those assumptions decide whether the agent is reliable.

Attio is one of the best-designed CRMs on the market, and it has leaned hard into AI. Relm makes a different bet: the operator is not a person with an assistant, it is the agent itself. That single assumption changes the API, the error design, the pricing and the UI. Here is the fair comparison.

Let us be clear up front: Attio is excellent. If you are choosing a CRM for a sales team of humans, its data model, speed, and design are genuinely top-tier, and its AI features are useful. This page is not a takedown. It is about a specific fork in the road - who does most of the reading and writing, a team or an agent - because that is where Attio and Relm diverge.

The same word, two philosophies

Attio is a UI-first CRM that added an API and AI. The center of gravity is the interface a team works in; the API is how you integrate around it. That is the right shape when people run the pipeline day to day.

Relm is an API-first CRM for LLMs. The center of gravity is the API and the native MCP server; the human dashboard is a thin window for the occasional look. That is the right shape when an agent runs the pipeline and a person only checks in.

Neither is "better." They are built for different operators, and everything below follows from that.

Four things an agent-first CRM does that a team CRM does not

Hands the agent the whole contract in one call. GET /v1/schema (or relm_describe_schema) returns every object, field, id prefix, enum value and list filter. The agent never scrapes docs or guesses your custom-field names.
Fails in a way the agent can fix. Every rejection is RFC-9457 problem+JSON with valid_options and a suggestion. Unknown stage? Here are the valid stages and the call to add one. The agent self-corrects in a turn - it never silently writes the wrong thing.
Is discoverable without an account. Public tools/list, an llms.txt, /.well-known manifests and an OpenAPI spec mean an agent (or a model deciding what to recommend) can learn Relm before anyone signs up.
Prices for bots, not seats. No per-user license for an agent that never logs into a UI - you pay per request.

Side by side

Dimension	Attio	Relm
Primary operator	A sales team (people), with AI assists	An AI agent, with a thin human view
Center of gravity	The UI	The API + MCP
Native MCP server	API + AI features; community MCP wrappers	First-party, 37 typed tools
A2A endpoint	-	Yes (JSON-RPC agent card)
Self-describing schema for LLMs	API metadata	`relm_describe_schema` - one live contract
Error contract	Standard API errors	RFC-9457 with `valid_options` + suggestion
Pricing basis	Per user seat	Per API request (no seats)
Free tier	Free plan for small teams	Free 1,000 req/mo, no card + free test mode

When to pick which

Pick Attio if a team of people is your primary operator, you want a best-in-class UI with AI assists, and per-seat pricing fits because those seats are humans doing real work in the interface. Pick Relm if the operator is an agent, you want the CRM to be fully drivable and discoverable over MCP / A2A / REST, and you would rather pay for the requests your automation makes than for seats no bot will ever use.

And it is not always a choice: keep a team CRM for the humans and let Relm be the CRM your triage, meeting-notes and follow-up agents own, wired together with webhooks. To try the agent-first side, give your AI agent a CRM in a few minutes.

FAQ

Is Relm an Attio alternative?

For the agent-run case, yes. Attio is a polished CRM for a team of people with AI assists; Relm is agent-first - the operator is the agent, the UI is thin, and everything is built to be driven over API and MCP. Teams of people fit Attio; an agent running the CRM fits Relm.

Does Attio have a native MCP server?

Attio has a strong REST API and AI features, with community MCP wrappers. Relm's MCP server is first-party - 37 typed tools plus relm_describe_schema - and there is an A2A endpoint and OpenAPI spec too. Agent access is the product, not an add-on.

How is pricing different?

Attio is per seat; Relm is per request, because the caller is an agent. Free is 1,000/mo (no card), Pro $29/mo for 100,000, Scale $249/mo for 2,000,000, plus a free test mode.

Can Relm and Attio be used together?

Yes - a human CRM for the team, Relm as the system your agents own, synced via each API and Relm's webhooks. Not always either/or.

Why choose an agent-first CRM at all?

Because a people-first CRM still assumes a human in the loop - it fails in ways a person would catch, documents itself for developers, and prices by seat. Agent-first inverts all three: one-call schema, machine-readable errors with valid_options, and no seat to buy for a bot.

Built for when the agent is the operator

Connect over MCP, discover the whole schema, and let your agent run the CRM. Free key, no card, free test mode.

Start free →

Relm vs Attio: who is the operator - a team, or an agent?