@irongraph/cli-darwin-arm64 0.1.7 → 0.1.9

package/README.md

@@ -1,128 +1,12 @@
 # IronGraph
 
-**A CPU graph database with built-in streaming and queues.**
+**Embeddable Cypher graph database with sub-µs point reads, built-in streams, queues, and local text embeddings. Standalone launcher.**
 
 Use Cypher to query graph structure, search text and vectors, inspect property history, and
 administer the database. Built-in Kafka-compatible Streams and AMQP-compatible Queues handle event
 and message workloads. The `irongraph` CLI starts and manages a standalone database with its
 built-in web console. IronGraph is open-source software released under the Apache License 2.0.
 
-## Performance at a glance
-
-Graph queries execute on the CPU against one shared canonical graph. Local text inference can
-use Metal or CUDA independently. Measure complete queries separately from individual graph operations;
-latency depends on the workload, graph size, indexes, and host.
-
-## Start the database and open the web console
-
-Use this path to run IronGraph as a standalone database with its built-in web console.
-Python, Node.js, and Rust embedding are separate options described below.
-
-### 1. Check the prerequisites
-
-You need Node.js 20.17 or later with npm, and an official `irongraph` release available on npm
-for your platform: macOS 15+ ARM64, or Linux ARM64/AMD64 with glibc 2.28+. Before the first
-official publication, these npm commands are not an available installation channel.
-
-Use a writable local data directory owned exclusively by this database process. First startup
-also needs network access and several gigabytes of free storage for the automatically installed
-local text embedding model. The launcher selects a compiled database binary with the web console
-included. You do not need Python, a Rust compiler, or a separate frontend server.
-
-### 2. Start IronGraph
-
-Run this command to install the matching package as needed and start in the background:
-
-```sh
-npx irongraph start --background
-```
-
-Your data lives in `~/.irongraph/data`. The database keeps running after you close the terminal.
-Wait for startup and the embedding model's installation, loading, and warm-up to finish. Check
-progress and readiness with:
-
-```sh
-npx irongraph logs
-npx irongraph status
-```
-
-Graph execution uses the CPU. Text inference selects its device independently; on a supported
-Mac, use `--embedding-backend metal`. On CPU-only hosts, including the published Linux packages,
-start with `npx irongraph start --background --embedding-backend cpu`. Automatic inference selection
-requires a supported accelerator. To run in the
-foreground with live terminal output, use `npx irongraph start` without `--background`.
-
-### 3. Open the web console
-
-When status reports **ready**, open **[http://127.0.0.1:18484/web/](http://127.0.0.1:18484/web/)**
-in a browser on the same computer.
-This address belongs to your running local instance. The product website at
-The public project page is not your database console.
-
-The console provides Query, Streams, Documents (for text documents), Training, Docs, and
-Settings. The graph visualization is the Plot result view inside Query.
-
-### 4. Create your first project
-
-Click **New** in the top bar, enter `notes`, and click **Create project**. The console selects
-your new project. In **Query**, replace the editor contents with:
-
-```cypher
-USE notes
-RETURN 'IronGraph is ready' AS status
-```
-
-Click **Run**, then select the **Table** result view. Expected result: a `status` column
-containing `IronGraph is ready`. Every graph operation needs
-an explicit project; IronGraph has no implicit default project. Next, open **Documents** to work
-with text documents in your project, or continue in Query to create nodes and relationships.
-
-### 5. Stop and restart
-
-For a background instance, run:
-
-```sh
-npx irongraph stop
-```
-
-For a foreground instance, press `Ctrl+C` and wait for the process to exit. Start again with
-`npx irongraph start --background` to reopen the same data directory. Your data persists between
-runs. Background mode does not install a system service or restart the database after a reboot.
-
-### Configuration and direct binary use
-
-Choose another directory with `--data-dir /absolute/path`. Pass the same directory to `status`,
-`logs`, and `stop`. A second start cannot open a second managed instance on that directory.
-
-Local listener defaults are consistent across development and releases:
-
-| Surface | Loopback port |
-| --- | --- |
-| Web console and Query API | `18484` |
-| Bolt | `18485` |
-| Kafka-compatible Streams | `18486` |
-| AMQP-compatible Queues | `18487` |
-| Local MCP HTTP | `18488` |
-
-The frontend development server uses `18489`; the standalone console is always served by the
-database's HTTP listener. Embedded libraries do not start these listeners.
-
-If the console does not open, check that the process is still running and startup completed.
-If port `18484` is already occupied, start with `--http-addr 127.0.0.1:19484` and open
-`http://127.0.0.1:19484/web/` instead. Keep plain listeners on loopback; access from another
-computer requires a configured remote listener with mutual TLS.
-
-You can also extract the matching official standalone archive and run its executable directly,
-without Node.js or npm:
-
-```sh
-./bin/irongraph --data-dir "$HOME/.irongraph/data" --http-addr 127.0.0.1:18484
-```
-
-Direct execution runs in the foreground. Stop it with `Ctrl+C`; launcher commands manage only
-instances started through the launcher. The archive also includes `bin/irongraph-mcp` for local
-stdio MCP hosts.
-
 ## Database capabilities
 
 | Capability | What you can do |
@@ -178,7 +62,7 @@ your release. CUDA requires a package built with CUDA support.
 ## Start with a text document in Python
 
 You need Python 3.9 or later, a matching official wheel, and a writable directory owned by one
-application process. Default startup installs, verifies, loads, and warms the local embedding
+application process. Default startup downloads, loads, and warms the local Bekko a8m embedding
 model; allow network access and space for the model on first use.
 
 After configuring your official package source, install the binary package: