
Alpha · public preview
One data platform. Every engine your app needs.
SQL, cache, streams, objects and search in one server, compatible with PostgreSQL, Redis, Kafka and S3. One identity, one backup, one console.
Engines
Built in, working together
KeelStore
Its own storage engine: copy-on-write B+trees, snapshots, checksummed pages, crash-safe commits.
CoreStore
ACID SQL with views, triggers, PL/pgSQL functions, window functions, vectors and geometry.
FlashCache
Every Redis data type, Lua scripts, pub/sub, blocking commands, eviction and persistence.
PulseStream
Partitioned streams, consumer groups, transactions, compaction and dead letters.
ObjectCore
Buckets, versions, multipart uploads, checksums and public links.
TrackSense
Ranked full-text search with highlights, kept in step with every change.
AutoGraph
Schema in code, relations, nested writes, migrations and generated types in four languages.
Migration Engine
Copies and follows PostgreSQL, MySQL, SQLite, Redis, Kafka and S3, verifies every row, then cuts over.
Control Center
A browser console for all of it, with live health for every engine.
Compatibility
Speaks the protocols you already use
PostgreSQL protocol
psql, node-postgres, Drizzle
Redis protocol
redis-cli, ioredis, node-redis
Kafka protocol
kafkajs, librdkafka
S3 API
AWS SDKs
Security
Safe by default
Row-level security by default
Every table starts closed. A tenant sees only its own rows, keys, messages, files and search hits.
Encrypted at rest
Row values and audit entries sealed with AES-256-GCM; TLS for HTTPS and every protocol.
Verified backups
Scheduled backups are checked before they count and carry a SHA-256 manifest.
Read replicas
A primary streams every commit to replicas, with synchronous commits, promotion and failover.
Signed releases
The installer checks each release's signature and rolls back an upgrade that fails its health check.
Measured, not claimed
Tested against the real systems
random SQL queries return the same results as PostgreSQL 18
Redis replies compared one by one with Redis 7.4
vector and geometry queries match pgvector and PostGIS
test suites run against the real systems Propus replaces
Questions
- What is Propus?
- Propus is one data platform with every engine an application needs: SQL compatible with PostgreSQL, a Redis-compatible cache, Kafka-compatible streams, S3-compatible object storage and full-text search, in one server with one identity, one backup and one console.
- Can I keep my existing drivers and tools?
- Yes. Propus speaks the PostgreSQL, Redis and Kafka protocols and the S3 API, so psql, node-postgres, Drizzle, redis-cli, ioredis, kafkajs, librdkafka and the AWS SDKs connect to it as they are. It also has SDKs for TypeScript, Python, Java and C#.
- Does Propus support pgvector and PostGIS?
- Yes. The vector type with HNSW indexes and the common PostGIS geometry and geography functions with spatial indexes are built in, and CREATE EXTENSION vector or postgis works as it does in PostgreSQL.
- Can I move an existing system into Propus?
- The Migration Engine copies data from PostgreSQL, MySQL, SQLite, Redis, Kafka and S3-compatible storage, follows changes while the application keeps running, verifies every row and then cuts over, with a rollback.
- Is Propus ready for production?
- Not yet. Propus is in alpha: every feature listed is built and tested against the system it replaces, and a public preview runs at propustech.tech. Contact us to discuss using it.
Try Propus
The preview runs at propustech.tech. The documentation covers installing, commands and every API.