SynapCores v1.13.0 — Your AI chat builds the first schema for you

The hardest part of adopting a new database is the blank page. You have to model the schema before you get any value. v1.13.0 removes that wall.

Describe your data, get a database

Open Get started, pick a template — support desk, product catalog, blog, analytics events — or just describe your own in plain English:

"a fleet-maintenance tracker: vehicles, service records, parts, and technicians"

The assistant creates the tables, adds a few sample rows, and shows you the schema and a first query. You're working in minutes, not hours.

Agentic scaffold chat — on by default

The AI Chat can now act, not just explain. Given a description it runs a non-destructive allowlist — CREATE TABLE, CREATE INDEX, INSERT — and a bundled CREATE … ; INSERT … executes each statement in order. Everything that could lose or rewrite data (DROP / DELETE / UPDATE / TRUNCATE / ALTER / GRANT) is always refused. It's the depth of a multi-model platform with the first-run simplicity of a toy. (Turn it off with AIDB_CHAT_AGENTIC_MODE=false.)

Installs that can't silently break

The Ubuntu-24.04 installer fix is carried in, and a new CI gate now runs the installer in clean ubuntu:22.04 and ubuntu:24.04 containers on every build — a package regression fails CI instead of reaching you. Version reporting is correct everywhere (/version, --version, the banner, the MySQL handshake).

Validated

feature_validator 88/88 · non-AVX-512 canary with no SIGILL · the agentic scaffold confirmed end-to-end on the published artifact (the agent created a table and inserted a row on the shipped native model) · recipe-cert 188/197, zero regressions.

Get it

curl -fsSL https://get.synapcores.com/install.sh | sh     # new install or in-place upgrade
docker pull synapcores/community:latest                    # = v1.13.0-ce

Re-running the installer upgrades in place and keeps your data and config. Linux x86_64 / aarch64 (Ubuntu 22.04 + 24.04) and macOS (Apple Silicon).

Running on a small box? See the System Requirements & Configuration guide — vector, ML and graph in ~2 GB.