bitsrun 0.0.0-stage → 0.5.42
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/README.md +41 -2
- package/dist/bitsrun.js +3542 -0
- package/package.json +43 -4
package/README.md
CHANGED
|
@@ -1,3 +1,42 @@
|
|
|
1
|
-
#
|
|
1
|
+
# bitsrun
|
|
2
2
|
|
|
3
|
-
|
|
3
|
+
Run open AI models as private, scale-to-zero APIs, billed per second. This is the BitsRun command line.
|
|
4
|
+
|
|
5
|
+
```
|
|
6
|
+
curl -fsSL https://get.bitsrun.com | sh # macOS and Linux
|
|
7
|
+
irm https://get.bitsrun.com/win | iex # Windows (PowerShell)
|
|
8
|
+
npm install -g bitsrun # or with Node 22+
|
|
9
|
+
|
|
10
|
+
bitsrun login
|
|
11
|
+
bitsrun deploy qwen3:8b --key
|
|
12
|
+
```
|
|
13
|
+
|
|
14
|
+
The install script picks the right binary for your computer, checks its SHA-256 before installing, and puts it in
|
|
15
|
+
`~/.local/bin` (Windows: `%LOCALAPPDATA%\bitsrun\bin`, added to your PATH). Set `BITSRUN_INSTALL_DIR` to choose
|
|
16
|
+
another folder. Update any time with `bitsrun upgrade`. Alpine needs `apk add libstdc++ libgcc` first.
|
|
17
|
+
|
|
18
|
+
## Sign in
|
|
19
|
+
`bitsrun login` opens bitsrun.com in your browser. Sign in, click **Allow**, and the terminal picks up the session.
|
|
20
|
+
Over SSH, or with `--no-browser`, it prints a link instead: open it anywhere, click Allow and paste the code it shows.
|
|
21
|
+
|
|
22
|
+
The session lasts 90 days and is saved in `~/.config/bitsrun/credentials.json` (readable only by you).
|
|
23
|
+
See and sign out every signed-in CLI under **API keys** in your dashboard, or run `bitsrun logout`.
|
|
24
|
+
|
|
25
|
+
## Commands
|
|
26
|
+
| Command | What it does |
|
|
27
|
+
|---|---|
|
|
28
|
+
| `bitsrun deploy <model> [--name] [--gpu l4\|l40s\|4xl4] [--idle 5m] [--key]` | Deploy an Ollama name (`qwen3:8b`), a Hugging Face repo (`hf:Qwen/Qwen3-8B` on vLLM, an embedding model on TEI, or a GGUF repo with an optional `:Q8_0`), an upload (`upload:<id>`), or a local `.gguf` file or `./model-folder` (uploaded first); `--key` also creates an API key for just this endpoint |
|
|
29
|
+
| `bitsrun hf-token [show \| set \| rm]` | The Hugging Face token for gated and private repos: `set` asks for it hidden, or reads it from a pipe (`echo "$HF_TOKEN" \| bitsrun hf-token set`); never an argument |
|
|
30
|
+
| `bitsrun upload <path>` | Upload a `.gguf` file or a safetensors model folder to deploy later (counts toward your plan's storage) |
|
|
31
|
+
| `bitsrun uploads [ls \| rm <upload> [--yes]]` | List uploads, or delete one no endpoint uses |
|
|
32
|
+
| `bitsrun ls` / `status <endpoint>` / `update <endpoint>` / `rm <endpoint>` | Manage endpoints, by name or id |
|
|
33
|
+
| `bitsrun update <endpoint> --always-on on\|off` | Keep one replica running (no cold starts), billed per second at the always-on rate; needs one day of credit to turn on |
|
|
34
|
+
| `bitsrun size <model>` | Which GPU a model needs, before you deploy |
|
|
35
|
+
| `bitsrun keys` / `keys create <name> [--endpoint]` / `keys revoke <key>` | API keys (secrets are shown once) |
|
|
36
|
+
| `bitsrun usage [--from] [--to]` | Running time and cost for a period |
|
|
37
|
+
| `bitsrun whoami` | Plan, balance and endpoint count |
|
|
38
|
+
| `bitsrun upgrade` | Update the CLI itself (standalone installs) |
|
|
39
|
+
|
|
40
|
+
Add `--json` to any command for machine-readable output. In CI, set `BITSRUN_TOKEN` to a CLI session token.
|
|
41
|
+
|
|
42
|
+
Help: support@bitsrun.com
|