Preview client SDKs for Python, TypeScript, Rust and Go plus six platform integrations, all Apache 2.0. They call the REST API that HeliosDB Nano and Lite serve; none is on PyPI, npm or crates.io yet, so you install them from source (the Go module with go get).
The SDKs and integrations in this catalogue call the HeliosDB REST API (/v1/...) that Nano and Lite serve; the Go SDK's own documentation limits it to those two editions. For HeliosDB Full, connect with a standard PostgreSQL driver: Full's HTTP gateway serves a smaller route set than the SDKs use. The PostgreSQL wire protocol is the surface all three editions share.
For AI agents, Code-Graph (hdb_code), Graph-RAG (WITH CONTEXT) and the native MCP server ship in HeliosDB Nano as build-from-source features; equivalents for Lite and Full are planned. See Code-Graph & Graph-RAG and CodeKB MCP.
Clients with type-safety, async and a fluent query builder, in preview. None is published to PyPI, npm or crates.io yet: the Go module installs with go get, and the others install from the HeliosDB-SDKs source.
Client with connection pooling and a query builder for REST and wire-protocol modes. Not published to PyPI — install it from source.
# not on PyPI - install from the SDKs repository
pip install "git+https://github.com/HeliosDatabase/HeliosDB-SDKs.git#subdirectory=sdks/python"A PEP 249 adapter shaped like the standard sqlite3 module, driving Nano underneath. Not published to PyPI — source only, and the supported subset is narrower than sqlite3. Read the scope before porting a codebase onto it. It is a client-side SDK adapter, not a database engine interface: the Release and Capability Matrix lists no SQLite engine for Nano.
# no release yet - build from source
git clone https://github.com/HeliosDatabase/HeliosDB-SDKs
pip install ./HeliosDB-SDKs/sdks/python-sqliteTyped client for Node.js, Deno, and Bun. Not published to npm — build it from source and install the local package.
# not on npm - build from the SDKs repository
git clone https://github.com/HeliosDatabase/HeliosDB-SDKs.git
(cd HeliosDB-SDKs/sdks/typescript && npm install && npm run build)
npm install ./HeliosDB-SDKs/sdks/typescriptHTTP client for remote HeliosDB servers, planned. For embedded use today, depend on the real, published heliosdb-nano crate directly — see the Nano quickstart.
# no release yet - build from source
git clone https://github.com/HeliosDatabase/HeliosDB-SDKsGo REST client for Nano (branches, vector search, time travel; agent memory planned for Nano) and Lite (SQL and branches; vector and agent-memory endpoints planned). The module is fetched straight from the SDKs repository, which tags its releases.
go get github.com/HeliosDatabase/HeliosDB-SDKs/sdks/go@latestIt is not a drop-in replacement for sqlite3, and it is not on PyPI. Here is what it does today, so you can decide whether it fits before you port anything onto it.
execute / executemany / executescript, fetchone / fetchmany / fetchall, and row objects with index- and name-based access.isolation_level=None) alongside explicit commit / rollback.?, :name and @name placeholder styles.create_function, create_aggregate, create_collation and load_extension raise NotSupportedError. User-defined functions, custom collations and extensions do not carry over.DatabaseError, not IntegrityError. Code that catches IntegrityError around inserts will not catch them here.heliosdb-nano REPL subprocess and parses its output. It is not an in-process binding, and it inherits that subprocess's lifetime and per-statement latency.The SQLite-compat adapter (import name heliosdb_sqlite) lives in sdks/python-sqlite and installs from a source checkout as the heliosdb-sqlite distribution. It needs a heliosdb-nano binary on PATH or in $HELIOSDB_BINARY. It is not published to PyPI.
If what you want is Nano inside a Python process rather than a sqlite3-shaped API, use the embedded binding. It is published on PyPI; the binding's releases trail the 4.41.1 engine.
pip install heliosdb-nano-embeddedThe import name is heliosdb_nano. This is a different thing from the adapter above — a direct embedded binding, not a DB-API shim over a subprocess. Starting fresh, start here.
heliosdb on the command lineNot published to npm — the name heliosdb is already taken by an unrelated package there, so this ships under a different name once released. Source only for now. For the engine itself today, use the real, published heliosdb-nano CLI — see the Nano quickstart.
# no release yet - build from source
git clone https://github.com/HeliosDatabase/HeliosDB-SDKs
# then, once built:
heliosdb start # persistent server
heliosdb start --memory # in-memory dev mode
heliosdb repl # interactive SQLDrop HeliosDB into the tools your team already uses.
SQL editor with intellisense, branch explorer, vector search panel, and natural-language query helper.
GitHub →Community node with query, vector, branch, and agent operations. Drop into any n8n workflow.
GitHub →Triggers and actions for workflow automation. Connect HeliosDB to 5 000+ Zapier-supported apps.
GitHub →Module definition for Make.com. Build no-code automations with HeliosDB as a data source / sink.
GitHub →REST API datasource configuration for building internal tools. Query HeliosDB from any Retool app.
GitHub →Multi-agent framework integration. HeliosDB as agent memory, retriever, and shared knowledge base.
GitHub →from heliosdb import HeliosDB
db = HeliosDB.connect("http://localhost:8080")
db.execute("CREATE TABLE docs (id SERIAL, body TEXT, embedding VECTOR(768))")
db.execute("INSERT INTO docs (body, embedding) VALUES ($1, $2)",
["Intro to HeliosDB", [0.1, 0.2, 0.3, ...]])
results = db.query(
"SELECT body FROM docs ORDER BY embedding <=> $1 LIMIT 5",
[query_vec],
)// package built from the HeliosDB-SDKs source (not on npm)
import { HeliosDB } from "@heliosdb/client";
const db = new HeliosDB({ url: "http://localhost:8080" });
await db.execute(
"CREATE TABLE docs (id SERIAL, body TEXT, embedding VECTOR(768))"
);
const results = await db.query(
"SELECT body FROM docs ORDER BY embedding <=> $1 LIMIT 5",
[queryVec],
);Every SDK is Apache 2.0 and targets the Nano and Lite REST API; for Full, use a PostgreSQL driver.