$ daptin-cli describe action integration install_integrationInspect an action’s instance requirements and input fields before executing it.
Open source · Self-hosted · LGPL-3.0
Your data model
is the backend.
Daptin reads table definitions from YAML or JSON and creates JSON:API routes for their records. It also supplies users, permissions, file uploads, server-side actions, integrations, and WebSocket events.
Go binary · SQLite, MySQL, or PostgreSQL · self-hosted
Tables:
- TableName: product
Columns:
- Name: title
DataType: varchar(200)
- Name: price
DataType: floatPOST /api/product 201
JSON:APIGraphQLOpenAPI
AuthEventsFiles
01 / RUN
A backend in one command
$ docker run --rm -p 127.0.0.1:6336:8080 daptin/daptin:v0.12.36
Open localhost:6336 →
CALL THE APIHTTP, CLI, and
HTTP, CLI, and
typed clients.
Call JSON:API routes directly, use the CLI, or install the JavaScript or Go client. All three can read records and run actions.
Explore developer tools →> await client.jsonApi.findAll<Product>('product')Use typed entities, JSON:API, actions, authentication, and direct storage uploads.
> new OpenAI({ baseURL: 'http://localhost:6336/v1' })Point the official client at Daptin and map model names to configured providers.
02 / BUILD
Define tables.
Define tables.
Daptin creates the API.
For each table, Daptin creates routes to read and change records, checks which users and groups may use them, publishes record changes, and lists the table in its web dashboard.
Read about tables and relationships →Database APIs
Create, read, update, and delete records through JSON:API. Filter, sort, paginate, include relationships, aggregate values, or enable GraphQL.
Explore APIs →Identity & permissions
Users, groups, ownership, JWT and OTP flows, plus table, row, relation, and action-level access control.
Read the permissions guide →Actions & schedules
Define input fields and ordered server-side steps, then call the action through HTTP or run it on a cron schedule.
Read the actions guide →Files & cloud storage
Attach files to records, save them locally or through rclone, and serve websites by host name or path.
Read the file storage guide →Realtime & protocols
Publish record changes over WebSockets, edit selected fields with YJS, and enable FTP, SMTP, or IMAP services.
Read the WebSocket guide →Integrations & LLMs
Import OpenAPI operations as actions and route OpenAI-compatible requests across configured LLM providers.
Read the integrations guide →
03 / SHIP
Three steps to
Three steps to
a running server.
- 1
Describe
Define tables, columns, relationships, validation, and table settings in YAML or JSON.
- 2
Configure
Create users and groups, set record permissions, and enable the actions, storage, integrations, and protocols you need.
- 3
Operate
Run one binary with health checks, TLS, audit logs, caching, clustering, and graceful shutdown.
RECORDS AND PERMISSIONS
Actions, uploads, and events
Actions, uploads, and events
use the same records.
Actions and file uploads refer to the same records exposed by the API. Record permissions also apply when Daptin runs an action or publishes an event.
Browse all server features →Your table definitions
Daptin server
APIsAuthActionsStorageRealtimeOps
CURRENT INTERFACES AND SOURCE
See the current interfaces
See the current interfaces
and source.
The metering UI below comes from the current demo repository. The schema, API routes, CLI commands, releases, and issue history are public.
TableName: product Columns: - Name: name DataType: varchar(200) # available after startup GET /api/product POST /api/product PATCH /api/product/{reference_id} # discover the running model daptin-cli list world daptin-cli describe action world import_data
RUNNABLE DEMOSRun the OAuth, LLM,
Run the OAuth, LLM,
metering, and integration flows.
Each repository includes setup instructions and exercises its flow against Daptin.
LLM API
Official OpenAI SDK compatibility
Chat, streaming, models, completions, embeddings, usage, and structured errors through Daptin’s /v1 routes.
OAuth 2.0 and OpenID Connect
Registers a client, signs in with authorization code and PKCE, exchanges the code for tokens, and calls the userinfo endpoint.
Inspect the demo ↗ API METERINGUsage, quotas, and credit accounting
Records API and LLM calls, rejects calls after a quota is reached, and debits a credit wallet after each metered call.
Inspect the demo ↗ STORED CREDENTIALSPer-user external API access
Stores OAuth and custom credentials for each user and verifies that one user cannot call an integration with another user’s credentials.
Inspect the demo ↗
04 / ADAPT
Common Daptin
Common Daptin
deployments.
Teams use Daptin as a primary application server or add a specific API, file, integration, metering, or realtime service to an existing stack.
01SaaS & multi-tenant appsUsers, groups, row permissions, plans, and quotas.→
02Internal tools & operationsStructured data, admin UI, actions, and integrations.→
03AI & integration gatewaysProvider routing, metering, credentials, and OpenAPI actions.→
04Realtime & collaborative productsEvents, WebSockets, feeds, and optional YJS.→
CLIENTS AND EXAMPLESClients, schemas,
Clients, schemas,
and application examples.
Install the CLI, use the JavaScript or Go clients, start from a schema package, or inspect an application repository.
Explore the ecosystem →
DEPLOYMENT
Run one Go binary
Run one Go binary
on your infrastructure.
Run Daptin as a single Go binary, choose SQLite for local work or MySQL/PostgreSQL for production, and keep your application data on infrastructure you control.
SQLite / MySQL / PostgreSQLLocal or cloud file storageDirect infrastructure billingHTTP / JavaScript / Go clients
Self-hosting and production operations →
RUN DAPTIN LOCALLY
Start with the