Platform

Purpose-built event store database for event sourcing

Append state changes to ordered streams, reconstruct state at a sequence or timestamp, replay accepted events into new projections, and resume durable consumers from acknowledged positions.

Start 14-day trial API docs Production patterns

Evaluate an event sourcing database on recovery

Start with one workflow. Append its accepted changes, rebuild a projection, restart the process, and compare recovered state with the expected result. Record persistence settings and failed cases alongside throughput.

Run the replay validation checklist Compare event stores and current-state databases Choose CQRS, event sourcing, or both

469K

events/sec

batch-ingest reference

11.9μs

p99 latency

Core indexed-read reference

WAL + Parquet

durable storage

checksummed persistence

default tenant MCP tools

73 with fleet controls

Database primitives for durable, queryable, and correct event history

Core event-store capabilities

Immutable Event Log

Accepted state changes are append-only events with CRC32 checksums. Existing event payloads are not updated in place, so their provenance remains queryable.

Time-Travel Queries

Reconstruct entity state at a historical timestamp by replaying accepted events. Core's 11.9us p99 indexed-read reference does not measure this reconstructive path.

WAL + Parquet Durability

A write-ahead log supports configurable fsync before events move to Parquet files with Snappy compression for analytical reads and compact storage.

DashMap Concurrent Reads

An in-memory concurrent map serves the published 11.9us p99 read benchmark. The separate batch-ingestion benchmark reaches 469K events/sec.

Projections & Snapshots

Build materialized views that stay in sync with your event stream. Checkpoint snapshots periodically to keep recovery fast as your event volume grows.

Schema Governance

Register event schemas and validate payloads at ingestion so incompatible changes are rejected before they enter a stream.

Ingest events and query history with plain HTTP

Write an event, then query its stream

Terminal

# Ingest an event
curl -X POST https://api.all-source.xyz/api/v1/events \
  -H "Authorization: Bearer $API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "event_type": "order.placed",
    "entity_id": "order-123",
    "payload": { "total": 149.99, "items": 3 }
  }'

# Time-travel query: what was this order's state yesterday?
curl "https://api.all-source.xyz/api/v1/events/query?\
  entity_id=order-123&before=2026-04-19T00:00:00Z" \
  -H "Authorization: Bearer $API_KEY"

Write your first event

Start with a 14-day trial on the hosted plans, or self-host the open-source core for free under Apache-2.0.

Start 14-day trial Compare with EventStoreDB