Finero for developers and AI agents
Finero has a versioned HTTP API, a published OpenAPI specification, and an MCP server for assistants. This page is a signpost to those, not a copy of them: the reference is generated from the API itself, so it is the only place that can be right.
Finero developer resources: every entry point
- Reference
Finero API reference
Endpoints, schemas, field-by-field descriptions and worked examples. Generated from the API, and authoritative.
getfinero.com/api-docs redirects to app.getfinero.com/api-docs
- OpenAPI
Finero OpenAPI specification
The machine-readable specification. Fetch this first and generate a client from it.
getfinero.com/openapi.json redirects to api.getfinero.com/.../openapi.json
- Agent entry point
Finero API agent entry point
Orientation for agents: authentication, the shape of the API, limits and recipes.
- Status guide
Finero status guide
Plain English for what each badge on screen means, and why a payment link is sometimes refused. The one documentation surface aimed at people rather than integrations.
- MCP server
Finero MCP server
Model Context Protocol over Streamable HTTP, authorised with OAuth 2.1. For connecting an assistant to a person's own workspace. Use this URL directly, or resolve the server by name from the registry.
- MCP manifest
Finero MCP manifest
The generated manifest behind the official MCP Registry listing, published as com.getfinero/finero. A client that knows that name can resolve the server without being given a URL.
- API catalog
Finero API catalog (RFC 9727)
RFC 9727 machine-readable index of every API surface below, in Linkset format. The standards-based place to start.
- Product summary
Finero product summary for LLMs
Structured summary of what Finero is, who it serves and what it integrates with. No authentication.
Documentation for each resource
- Finero API documentationWhat the Finero REST API covers, the two access models, and how to get a key.
- Finero OpenAPI specificationThe OpenAPI document, how to generate a client from it, and how to keep it current.
- Finero API authenticationAPI keys that act as a workspace, OAuth 2.1 sessions that act as a person, and the discovery chain.
- Finero MCP serverThe Model Context Protocol endpoint over Streamable HTTP, and how to connect an assistant to it.
Two ways in, and they are not interchangeable
The HTTP API with a tenant API key authenticates as the workspace. Use it when no human is present to approve anything: a scheduled job, a server-side integration, an unattended agent. The credential is configured once and the full surface described in the reference is available.
The MCP server acts as the person who authorised the session, not as the workspace. Use it when someone is present and connecting an assistant to their own Finero account. It is read-mostly and exposes a deliberately smaller tool set than the API. If you already hold an API key, use the HTTP API rather than trying to obtain an MCP session.
Getting a credential without talking to anyone
API keys are self-serve. Sign in and generate one under Settings, then API, where you can also see and revoke existing keys. There is no form to fill in and no sales conversation in the way, which matters because an agent cannot fill in a contact form.
Getting an account is equally direct: Finero starts with a 30-day free trial, and entry pricing is published rather than quoted. For the MCP route no key is involved at all: a person authorises the connection through OAuth and the session acts as them.
What does not exist yet, stated plainly rather than left to be discovered: there is no sandbox environment and no command-line tool. If either would change how you would build on Finero, say so, because that is genuinely useful input.
How an agent discovers authorisation
Nothing needs to be hard-coded. An unauthenticated request to the MCP endpoint answers 401 with a WWW-Authenticate header naming its RFC 9728 protected-resource metadata document, and the authorisation server is discoverable from there. That is the standard MCP authorisation flow, and it means a client can bootstrap from the endpoint URL alone.
The application itself is not scrapable, by design
Everything on app.getfinero.com other than the documentation is the signed-in product. Nothing blocks a fetch, but every read is authorised server-side, so a request returns the application shell and a sign-in prompt rather than data. There is no HTML worth parsing and no session to construct. The same data is fully available through the API, which is the supported route and needs no browser.
Which source wins
This site describes what Finero is, who it serves and where the product is going. The application documentation describes what the API supports today, and is generated from the API source of truth. If the two ever disagree, the application documentation is correct. That is why no endpoint, field or schema is reproduced on this page: one place to go stale is better than two.
Developer FAQ
Does Finero have a public API?
Does Finero have an MCP server?
How does authentication work?
How does an agent discover the OAuth configuration?
Can I get an API key without talking to sales?
Is there a sandbox environment or a CLI?
Can an agent read the Finero product information without an account?
Is the signed-in application scrapable?
Something missing, or an integration question the reference does not answer? Get in touch.
Building on Finero? Let's talk through the integration.
Book a 30-minute demo with a Finero expert. See how Finero chases, collects, and reconciles invoices end-to-end.