Query across data models.
Work with relational data, vector similarity, and graph relationships without maintaining a separate store for each.
Explore the architecture ↗THE SELF-HOSTED DATABASE FOR AI
Build AI apps
with fewer
moving parts.
SQL, vector search, graphs, and durable agents in one database. Bring your AI closer to your data, on infrastructure you control.
Free Community Edition · No signup or credit card
Community Edition v2.1.3 ↗macOS & Linux · Windows via WSL2THE ARCHITECTURE01 / ONE ENGINE
Your AI application
SQL / REST / MCP
SynapCoresOne database. Connected capabilities.
≡SQL⋮Vectors⌘Graphs↗Agents◴Memory✳Inference
Your laptop. Your server. Your data.
LESS TO CONNECT. MORE TO BUILD.
Your data stack should
move your application forward.
Every separate service brings another connection, deployment, and failure mode. SynapCores puts related capabilities in the same engine.
Give your agents a home.
Declare durable agents in SQL, trigger them from data changes, and retain memory and run history in the database.
Build your first agent ↗Choose where AI runs.
Use embedded local inference, or configure an external model provider. Choose the deployment that fits your data requirements.
Check system requirements ↗THE PRODUCT, IN PRACTICE
Follow the answer
back to the data.
A support agent remembers a conversation, retrieves context, and selects a tool. The debug panel makes the database’s role visible.
Inspect the reference application ↗BUILD SOMETHING YOU CAN INSPECT
Start with the code, follow the data, then reproduce the workflow on your own installation.
Explore the agent lifecycle in SQL
See it fire
INSERT a row. Watch the agent act.
Three statements, one engine. No service to deploy, no queue to drain, no worker to page.
Watch a real agent run end-to-end →
The same loop, recorded live against the engine — no mockups.
A CLEAR FIRST STEP
From install
to your first query.
Start with a local evaluation. Explore the bundled interface, then bring your own application.
- 01
Install Community Edition.
Choose your platform and review the system requirements.
Get the install command ↗ - 02
Open the bundled web UI.
Use the URL and first-boot credentials printed by the installer.
Follow the quickstart ↗ - 03
Run a working example.
Choose a recipe, execute the query, and inspect the result.
Choose your first recipe ↗
A WORKING STARTING POINT
Your next idea.
Already started.
Runnable SQL and Cypher, with the steps to make them your own. Explore 246 recipes across agents, graph, ML, and more.
AGENT MEMORY12 min
Give Any Agent Long-Term Memory
Store and recall memories by meaning with vector embeddings in plain SQL. Works with Claude Code, OpenClaw, LangChain, or any custom agent.
Open recipeAGENT MEMORY12 min
Recall Memory by Meaning + Recency
Blend semantic relevance with recency so the agent surfaces the right memory now, not a stale one — one SQL query, no extra cache layer.
Open recipeAGENT MEMORY14 min
Episodic + Semantic Memory
Two memory types — episodic (what happened) and semantic (durable facts) — in one SQL store. Automatically promotes repeated episodes to facts.
Open recipeEXPLORE MORE WORKFLOWS
CHOOSE YOUR STARTING POINT
One engine, when you want
less infrastructure to assemble.
Different architectures suit different workloads. Evaluate the tradeoffs with your own data.
| Approach | A good fit when… | What to evaluate |
|---|---|---|
| Postgres + extensions Your existing foundation | Your application already depends on the Postgres ecosystem. | How extensions and external AI services fit your workflow. |
| Specialist services A composed stack | You need dedicated services for retrieval, graphs, or orchestration. | Integration effort, operations, and independent scaling. |
| SynapCores One self-hosted engine | You want SQL, vectors, graphs, and durable agents together. | Workload fit, compatibility, and edition requirements. |
START WITH COMMUNITY EDITION
Put the database
to work. For free.
Evaluate on your own machine with the core database and AI capabilities. No sales call to get started.
$0 Community Edition
Free under the Community Edition license.
Infrastructure and optional external model usage are separate.
Included in Community Edition
- SQL, vector search, and graph queries
- Durable agents and persistent memory
- Embedded inference and model integrations
- Bundled web UI and example recipes
- REST API and native MCP access
EVALUATING FOR YOUR TEAM?
Plan your production deployment.
Review Enterprise options for clustering, advanced access controls, and SSO. Paid production support is also available for Community Edition.
Compare editions ↗Talk to engineering ↗Start with Community Edition ↗Compare all Community and Enterprise features
What's in Community Edition
Community Edition is a complete AI-native engine. Sales is an upgrade path, not a front door.
| Feature | Community | Enterprise |
|---|---|---|
| Core SQL Engine | ||
| Vector Search (HNSW) | ||
| Graph Database (Cypher) | ||
| AI / LLM Integration | ||
| MCP Server (Model Context Protocol) | ||
| Multimedia (PDF/AV) | ||
| AutoML | ||
| Multi-node Clustering | ||
| Raft Replication | ||
| Fine-grained RBAC | ||
| Audit Logging (Scale) | ||
| SSO / SAML / LDAP | ||
| Immutable Tables |
New in v1.11: SynapCores now speaks the MySQL binary wire protocol — point your existing MySQL driver, ORM, or CLI at it for a near drop-in replacement, no rewrite.
Is the Community Edition really free?
Yes. It is available without signup or a credit card, under the Community Edition license. It is proprietary software; free availability does not mean an open-source license. Your hosting costs and any external AI provider charges are separate. Review the license information.
Does my data leave my infrastructure?
Local storage and embedded inference can run on your own infrastructure. If you configure an external AI provider, requests to that provider can include your data. Review the model configuration for your workload before using sensitive data.
Do I need a GPU?
A GPU is not required for CPU inference. Model choice and workload affect memory needs and response time. Check the current requirements for your system.
Can I replace MySQL or Postgres without changes?
Plan a compatibility evaluation. SynapCores supports the MySQL wire protocol, but that alone does not guarantee support for every SQL statement, ORM behavior, or extension. Test your schema, queries, transactions, and tools. Explore the engine.
What is the best way to evaluate it?
Install locally, run a recipe, and reproduce a small representative workload. Then evaluate correctness, latency, recovery, and the controls you need before choosing a production deployment. Open the quickstart.
BUILD WITH US
A real workload.
A direct line to engineering.
Help shape SynapCores as a design partner, with priority on the features your team needs.
Become a design partner ↗YOUR NEXT APPLICATION STARTS HERE
Less setup.
More possibility.
Install free
Community Edition · No signup