# Lium GPU Cloud (pypi · lium.io)

Rent NVIDIA GPU pods on Lium, run commands on them and remove them, billed per second.

- Trust score: 77/100 (medium)
- Registry status: active
- Liveness: live
- Owner verified: no
- Last scored: 2026-10-23

## Components

- pypi · `lium.io`: 77/100 (this document), [markdown](https://verifymcp.io/servers/datura-ai-lium/lium-io.md), [page](https://verifymcp.io/servers/datura-ai-lium/lium-io)

## Channel facts

- Registry: `pypi`
- Package: `lium.io`
- Version: `0.28.2`
- Transport: `stdio`

## Trust breakdown

How this component scores in each security and reliability category. Every signal is checked automatically from public evidence about the published package, including repeated runs of it in an isolated sandbox, and we only credit what we can confirm. Scores are 0–100 per category. Scoring method: https://verifymcp.io/docs/scoring (what has changed: https://verifymcp.io/docs/scoring/changelog)

Scored 2026-10-23.

- **Supply Chain Security**: 99/100
  - No malware found by supply-chain analysis.
  - Known CVEs were checked across the 49 of 50 dependencies we could resolve, so this covers what we could see, not the whole tree.
  - Runs hatchling.build at install time, a recognised build step with no custom scripting around it.
  - Dependency health was assessed across the 49 of 50 dependencies we could resolve, so this covers what we could see, not the whole tree.
- **Provenance & Transparency**: 97/100
  - Source repository is publicly reachable at the declared URL.
  - Cryptographically verified build provenance (signed, bound to Datura-ai/lium).
  - Clear OSI-approved license (MIT).
  - Actively maintained (last published 0 days ago).
  - Disclosure check failed: no security disclosure policy was found in the source repository.
- **Schema Quality & AI Usability**: 71/100
  - AI-judged instruction clarity (good).
  - Tool/resource definitions use about 653 tokens (~108/item across 6 items; 6 tools + 0 resources), lean.
  - Usage-examples check failed: none of the tools include examples.
- **Stability & Change Management**: 10/100
  - Stability observed for 3 of 30 days with no destabilising changes; credit accrues until the full window elapses.
- **Tool Coverage**: 71/100
  - 100% of tools have a non-trivial description (not blank, and not just the tool's name).
  - 0% of tool parameters carry a description.
  - Structured output schemas are declared (100% of tools); any adoption earns full credit.
- **Tool Safety**: 75/100
  - No prompt-injection markers were found in the server instructions, tool names or descriptions we captured.
  - 0 of 1 tool(s) whose name or description implies an irreversible operation declare an MCP destructiveHint annotation; "remove_pod" implies "remove" and declares no destructiveHint at all, which the MCP spec reads as destructive by default.
  - An AI judge read all 7 captured unit(s) of tool text and found none that tries to manipulate the model reading it.
- **Capabilities**: 100/100
  - Implements a current MCP spec version (2026-07-28).

## Install

### How do I install the Lium GPU Cloud MCP server?

Lium GPU Cloud runs locally as a PyPI package, launched with uvx lium.io. Ready-made configuration for Claude, Cursor, VS Code, Codex and 5 more is on this page, copied from each client's own documentation.

### Claude

```bash
claude mcp add datura-ai-lium -- uvx lium.io
```

### Cursor

```json
{
  "mcpServers": {
    "datura-ai-lium": {
      "command": "uvx",
      "args": [
        "lium.io"
      ]
    }
  }
}
```

### VS Code

```json
{
  "servers": {
    "datura-ai-lium": {
      "command": "uvx",
      "args": [
        "lium.io"
      ]
    }
  }
}
```

### Codex

```bash
codex mcp add datura-ai-lium -- uvx lium.io
```

### opencode

```json
{
  "$schema": "https://opencode.ai/config.json",
  "mcp": {
    "datura-ai-lium": {
      "type": "local",
      "command": [
        "uvx",
        "lium.io"
      ],
      "enabled": true
    }
  }
}
```

### OpenClaw

```bash
openclaw mcp add datura-ai-lium --command uvx --arg lium.io
```

### Hermes

```yaml
mcp_servers:
  datura-ai-lium:
    command: "uvx"
    args: ["lium.io"]
```

### Netclaw

```json
{
  "McpServers": {
    "datura-ai-lium": {
      "Transport": "stdio",
      "Command": "uvx",
      "Arguments": [
        "lium.io"
      ]
    }
  }
}
```

### Vellum

```bash
assistant mcp add datura-ai-lium -t stdio -c uvx -a lium.io
```

### Other

```json
{
  "mcpServers": {
    "datura-ai-lium": {
      "command": "uvx",
      "args": [
        "lium.io"
      ]
    }
  }
}
```

## Changelog

Every change recorded for this component, newest first. Days that predate change tracking, or that we cannot explain, say so: "we were watching and nothing happened" and "we were not watching" are different claims.

### 2026-10-23 (score 77, 0)

- [functional] Package version: 0.28.0 → 0.28.1

### 2026-10-22 (score 77, +1)

- [security regression] Malware scan: pass → unverified
- [functional] Package version: 0.26.0 → 0.28.0
- [functional] Package version: 0.26.0 → 0.27.2

### 2026-10-21 (score 76, +15)

- [security improvement] Malware scan: unverified → pass
- [functional improvement] Stability: unverified → 0.03
- [functional] Package version: 0.25.0 → 0.26.0

### 2026-10-20 (score 61)

First indexed and scored.

## MCP tools (6)

### `list_gpus` (~127 tokens)

Use this when you need a GPU: lists Lium nodes for rent, cheapest first, with the executor_id rent_pod
takes, tier ("secure", or "spot": cheaper, may be reclaimed), reliability_score (0-100) and
rent_price_per_hour_usd (what rent_pod bills per hour for the node's free GPUs).
gpu_type filters by model ("H100", "RTX 4090", "B200"); gpu_count by GPUs on the node.

Input parameters:

- `gpu_count`
- `gpu_type`
- `limit` (integer)

Output parameters:

- `result` (object)

### `rent_pod` (~277 tokens)

Use this when you have picked a node and want to rent a GPU pod on it: rents it and waits until it is
running. executor_id is the executor_id from list_gpus, or the id (or rent_url) that docs.lium.io/mcp
search_gpus returns. With dry_run=true it rents nothing and returns a quote: tier, reliability, cost per
hour and your balance; show that to the user before renting. This spends money: the pod is billed per
second at the node's hourly price until remove_pod. Give it a short unique name. If the result has status "pending", the pod is already rented: do
not rent again, poll list_pods or remove_pod its pod_id.
If it has status "recovered", nothing new was rented: it lists pods earlier calls rented that you have not
used yet. To rent another pod next to those, pass their ids as known_pod_ids. One rent_pod runs at a time:
while one is still running, another returns the error rent_in_progress and rents nothing.

Input parameters:

- `dry_run` (boolean)
- `executor_id` (string, required)
- `known_pod_ids`
- `name` (string, required)
- `template_id`
- `timeout_seconds` (integer)

Output parameters:

- `result` (object)

### `list_pods` (~31 tokens)

Use this when you need your pods: lists them with status, GPU, price per hour and SSH command.

Output parameters:

- `result` (object)

### `run_command` (~76 tokens)

Use this when you want to run something on a pod: runs a shell command on a running pod (by pod id, or a
huid or name that matches one pod) and returns exit code, stdout and stderr.

Input parameters:

- `command` (string, required)
- `pod` (string, required)
- `timeout_seconds` (integer)

Output parameters:

- `result` (object)

### `remove_pod` (~53 tokens)

Use this when you are done with a pod: removes it (by pod id, or a huid or name that matches one pod).
Billing stops; anything on its disk is gone.

Input parameters:

- `pod` (string, required)

Output parameters:

- `result` (object)

### `get_balance` (~30 tokens)

Use this when you need to know whether you can afford a rental: returns your Lium balance in USD.

Output parameters:

- `result` (object)

## Diagnostics

Captured diagnostic sections: Provenance, Install scripts, Dependencies. The full working is on the page: https://verifymcp.io/servers/datura-ai-lium/lium-io#diagnostics

## Score history

- 2026-10-23: 77
- 2026-10-22: 77
- 2026-10-21: 76
- 2026-10-20: 61

## Common questions

### What is the Lium GPU Cloud MCP server?

Lium GPU Cloud is an MCP server listed in the public MCP registry as io.github.Datura-ai/lium. Rent NVIDIA GPU pods on Lium, run commands on them and remove them, billed per second. This page covers its PyPI package (lium.io).

### Is the Lium GPU Cloud MCP server safe to use?

Lium GPU Cloud scores 77 out of 100 on VerifyMCP. Its build provenance is signed and verified. That is a record of what we were able to check automatically, not an endorsement. The category breakdown on this page shows every signal behind the number, including the ones we could not confirm.

### What tools does the Lium GPU Cloud MCP server expose?

Lium GPU Cloud exposes 6 tools: list_gpus, rent_pod, list_pods, run_command, remove_pod, get_balance. Their descriptions and schemas cost roughly 594 tokens of context every time the server is loaded.

### Is the Lium GPU Cloud MCP server still maintained?

Lium GPU Cloud is still listed as active in the MCP registry. We last reached this channel on 23 October 2026. Those dates come from our own scans of the registry and the channel itself, not from anything the publisher announced.

### What licence is the Lium GPU Cloud MCP server under?

Lium GPU Cloud declares the MIT licence, which is OSI-approved. That covers the source only, and says nothing about the cost of any service it calls.

## Links

- PyPI project: https://pypi.org/project/lium.io/
- Socket report: https://socket.dev/pypi/package/lium.io
- Repository: https://github.com/Datura-ai/lium
- Website: https://lium.io/
- Changelog RSS feed: https://verifymcp.io/servers/datura-ai-lium/lium-io.xml
- Changelog JSON feed: https://verifymcp.io/servers/datura-ai-lium/lium-io.json
- HTML version of this page: https://verifymcp.io/servers/datura-ai-lium/lium-io