What is a metadata-driven CRM?

A metadata-driven CRM stores fields, relations, views, and permissions as data. One runtime interprets the definitions to render every product surface — so admins change the system through configuration, not code.

Can I add custom fields without code?

Yes. Fields, relations, and views are metadata definitions — configuration changes, not engineering changes. Both the UI and the REST API can manage them.

Is BeeCRM available today?

Phase 1 (CRM spine) is in active development. Early access signups are open. There is no publicly committed general-availability date.

Every UI operation — including metadata and admin operations — has a REST equivalent with 100% UI-operation parity. API tokens are attenuated sub-principals of a real user.

Is BeeCRM built for agencies and service businesses?

Yes. The lead-to-customer-to-task loop, custom fields, role-based permissions, and kanban views are designed for service and agency businesses with 10–200 staff.

How is BeeCRM different from legacy CRM suites?

Legacy SMB CRM suites have fixed schemas — adding a field or a view requires engineering work. BeeCRM stores that configuration as metadata, so your own admins extend the system without a migration project.

Phase 1 · in developmentEarly access — opening soon.

Your CRM, defined as data.

BeeCRM is a metadata-driven CRM platform. Fields, relations, views, permissions, and workflow are stored as data and interpreted by one shared runtime — so your own admins extend the system without an engineering project.

Join early access Read the specs

Definitions

Object · Lead

fields: typed
views: defined
actions: governed

Field · service_region

type: select
required: false
scope: workspace

Permission · Lead

role × object × action

Metadata engine

resolve schema
derive field state
compile permissions
render view
expose operation

one source of behavior

Surfaces

List + filter

Detail + relations

Edit + validation

Kanban + bulk

REST API parity

same permission boundary

Definitions as data → one runtime → every product surface

Extension should not mean rebuilding.

In most CRMs, every new field, relation, or view is a small engineering project — a fixed schema, a consultant, a migration. BeeCRM makes those changes configuration, while one rendering and authorization model stays intact.

How it works

One definition. Every surface.

Object and field definitions enter one metadata engine. The same definitions shape the interface, enforce permissions, and expose equivalent operations over REST.

Definitions

Object · Lead

fields: typed
views: defined
actions: governed

Field · service_region

type: select
required: false
scope: workspace

Permission · Lead

role × object × action

Metadata engine

resolve schema
derive field state
compile permissions
render view
expose operation

one source of behavior

Surfaces

List + filter

Detail + relations

Edit + validation

Kanban + bulk

REST API paritysame permission boundary

Definitions
Shared runtime
Rendered surfaces

The API can do what the UI can do.

UI operations — including metadata and admin operations — have 100% UI-operation parity over REST. API tokens are attenuated sub-principals of a real user, so automation inherits the user's bounds instead of escaping them.

An agent never outranks its human.

POST /api/v1/objects/{type}/fields

POST /api/v1/objects/lead/fields
Authorization: Bearer <user-token>
Content-Type: application/json

{
  "key": "service_region",
  "label": "Service region",
  "type": "select",
  "required": false
}

Leads become customers. Work stays visible.

Leads, customers, contacts, and tasks share custom fields, role × object permissions, audit, bulk operations, and metadata-defined views in the Phase 1 CRM spine.

BeeCRM Leads kanban view in light theme — **Leads pipeline**Metadata-defined stages, bulk-aware movement — from first touch to converted customer.

BeeCRM Leads kanban view in dark theme — **Leads pipeline**Metadata-defined stages, bulk-aware movement — from first touch to converted customer.

BeeCRM customer detail view in light theme — **Customer detail**Relations, activity, tasks, and custom fields in one record — all metadata-defined.

BeeCRM customer detail view in dark theme — **Customer detail**Relations, activity, tasks, and custom fields in one record — all metadata-defined.

BeeCRM Tasks kanban view in light theme — **Tasks board**Owner, due state, and permission-aware work queues — same metadata drives every view.

BeeCRM Tasks kanban view in dark theme — **Tasks board**Owner, due state, and permission-aware work queues — same metadata drives every view.

Built for the teams who outgrow fixed CRMs.

Service & agency businesses

10–200 staff

Replace fixed-schema tools with a CRM your own admins extend. Lead → customer → task loop, custom fields, kanban views — all configuration, no engineering project.

Platform & RevOps teams

Governance from day one

Role × object permissions, field-level modes, audit, team scopes. A metadata-driven core you can govern and extend without waiting on a vendor.

Developers & AI integrators

Full API parity

Every UI operation has a REST equivalent. Attenuated tokens mean an AI agent can drive the entire product — but never with more authority than the human it acts for.

Join early access

Forward-looking targets

A core that admits more modules.

CRM establishes the shared runtime. Sales documents, engagement, and the bee CLI follow next; projects and billing beyond. Planning windows, not promises — phases exit on proof, not on dates.

Phase 1

2026 – early 2027

CRM

Metadata engine, custom fields, leads → customers → tasks, RBAC, API tokens. Active development.

Phase 2

2027

Sales docs Engagement bee CLI

Estimate → invoice, recurring run reconcile, bee CLI at 100% REST parity.

Phase 3

2027 – 2028

Projects Billing

Hierarchy + status-transition governance, advanced business modules.

Sprint-aligned planning targets on 2–3 week cadence. Targets are revisable; exit criteria are not. See full roadmap for milestone details.

Portable by design. Permission-first by default.

TypeScript end-to-end. Hono + PostgreSQL on Cloudflare Workers, portable to a conventional Node server. Designed to remain operable by a small team.

RuntimeEdge + server portable
CoreTypeScript end-to-end
ClientReact + Tailwind, metadata-rendered
TransparencyRead the specs

Questions

What is a metadata-driven CRM?: A metadata-driven CRM stores fields, relations, views, and permissions as data. One runtime interprets the definitions to render every product surface — so admins change the system through configuration, not code.
Can I add custom fields without code?: Yes. Fields, relations, and views are metadata definitions — configuration changes, not engineering changes. Both the UI and the REST API can manage them.
Is BeeCRM available today?: Phase 1 (CRM spine) is in active development. Early access signups are open. There is no publicly committed general-availability date.
Is there an API?: Every UI operation — including metadata and admin operations — has a REST equivalent with 100% UI-operation parity. API tokens are attenuated sub-principals of a real user.
Is BeeCRM built for agencies and service businesses?: Yes. The lead-to-customer-to-task loop, custom fields, role-based permissions, and kanban views are designed for service and agency businesses with 10–200 staff.
How is BeeCRM different from legacy CRM suites?: Legacy SMB CRM suites have fixed schemas — adding a field or a view requires engineering work. BeeCRM stores that configuration as metadata, so your own admins extend the system without a migration project.

Follow the build. Join when it is ready.

BeeCRM is in active development. Leave one email address and we will only use it to contact you about early access.