Platform capabilities
Everything AltBase ships in one backend platform
These sections map directly to the workspace modules: `atlas-api-engine`, `atlas-auth`, `atlas-storage`, `atlas-realtime`, `atlas-workflows`, `atlas-vector`, and the operator-facing apps and tools around them.
Module 1 of 8
Database APIs
PostgREST-compatible REST CRUD, raw SQL execution, schema introspection, two-tier query cache, and an index advisor — all over tenant-isolated PostgreSQL schemas.
Code example
const { data: invoices } = await altbase
.from("invoices")
.select("id,total_cents,status")
.eq("status", "open");
const { data: health } = await altbase.sql("select now() as server_time");Key specs
- PostgREST-compatible REST CRUD
- Raw SQL execution endpoint
- Two-tier query cache (Redis + in-process)
- Automatic index advisor
- Schema introspection API
- Tenant-isolated PostgreSQL schemas
Module 2 of 8
Auth & SSO
Email, magic link, OTP, OAuth, OIDC, and SAML with RS256 JWTs, refresh-token families, MFA/TOTP, 47 enterprise SSO presets, B2B customer organizations, and row-level security.
Code example
await altbase.auth.signUp({
email: "owner@acme.dev",
password: "correct-horse-battery-staple",
options: {
data: { role: "owner" },
},
});Key specs
- 47 SSO presets out of the box
- RS256 JWT tokens with MFA/TOTP
- SAML 2.0 support
- OIDC-compliant flows
- Refresh-token families with theft detection
- B2B customer organizations
Module 3 of 8
Object Storage
Buckets, signed URLs, resumable uploads via TUS v1.0.0, multipart flows, image transforms, and pluggable backends for local filesystem, S3, and Azure Blob.
Code example
await altbase.storage.from("documents").upload(
"contracts/q1.pdf",
file,
{ contentType: "application/pdf" },
);Key specs
- TUS v1.0.0 resumable uploads
- 3 storage backends (local, S3, Azure Blob)
- Image transforms on the fly
- Signed URLs with expiration
- Multipart upload flows
- Per-bucket access policies
Module 4 of 8
Realtime
WebSocket subscriptions with CDC-driven delivery under 50 ms, presence tracking, client-to-client broadcast, and NATS JetStream fan-out with RLS enforcement.
Code example
altbase
.channel("live-orders")
.on("postgres_changes", {
event: "*",
schema: "public",
table: "orders",
}, payload => console.log(payload));Key specs
- Sub-50 ms CDC-driven delivery
- WebSocket subscriptions
- Presence tracking
- Client-to-client broadcast
- NATS JetStream fan-out
- RLS enforcement on channels
Module 5 of 8
Functions & Triggers
Multiple runtimes: QuickJS for sub-10 ms in-process execution with host APIs, Wasmtime for sandboxed Wasm with fuel metering, plus Node and Deno support. Unified triggers from CDC, auth, storage, cron, and custom events.
Code example
await fetch("http://localhost:4000/functions/v1/invoke/refresh-search", {
method: "POST",
headers: { Authorization: "Bearer " + serviceKey },
});Key specs
- QuickJS: sub-10 ms in-process runtime
- Wasmtime: sandboxed Wasm isolation with fuel metering
- Multiple runtimes: QuickJS, Wasmtime, Node, Deno
- CDC, auth, storage, cron triggers
- Host API bindings (db, http, kv, storage)
- Unified trigger model with condition evaluation
Module 6 of 8
Workflows & Integrations
DAG-based workflow engine with HTTP, function, SQL, delay, approval, branch, loop, parallel, and error-handler steps. 500+ SaaS integrations via Nango with webhooks and circuit breakers.
Code example
await fetch("http://localhost:4000/workflows/v1/runs", {
method: "POST",
headers: { Authorization: "Bearer " + serviceKey },
body: JSON.stringify({ definition_id: "wf_sync_accounts", input: { source: "hubspot" } }),
});Key specs
- DAG-based workflow engine
- HTTP, SQL, function, delay steps
- Approval, branch, loop, parallel nodes
- 500+ SaaS integrations via Nango
- Webhook ingress with circuit breakers
- Error-handler steps
Module 7 of 8
Vector Search & RAG
Embedding generation via OpenAI, Azure, Ollama, or local ONNX. Hybrid search with RRF fusion, four vector backends, knowledge collections with chunking, and an ask-with-citations endpoint.
Code example
await fetch("http://localhost:4000/v1/projects/proj_123/knowledge/support/ask", {
method: "POST",
headers: { Authorization: "Bearer " + serviceKey },
body: JSON.stringify({ question: "How do we rotate API keys?" }),
});Key specs
- OpenAI, Azure, Ollama, ONNX embeddings
- Hybrid search with RRF fusion
- 4 vector storage backends
- Knowledge collections with chunking
- Ask-with-citations endpoint
- Configurable chunk strategies
Module 8 of 8
GraphQL
Auto-generated schema from Postgres with queries, mutations, and CDC-driven subscriptions. Redis-cached schemas, DataLoader batching, and a built-in GraphiQL IDE.
Code example
query PipelineSummary {
opportunities {
nodes {
id
stage
amount
}
}
}Key specs
- Auto-generated schema from Postgres
- Queries, mutations, and subscriptions
- CDC-driven live subscriptions
- Redis-cached schema compilation
- DataLoader batching
- Built-in GraphiQL IDE