Fine-tune Qwen models with LoRA, export to GGUF, and manage local LLM deployments through a terminal interface or MCP server.
Also known as ondeinference/onde-cli, onde-cli, @ondeinference/cli
- Stars
- 5
- Tools
- —
- Downloads, 30 days
- —
- Latest version
- 0.5.1
- Last commit
- Nov 10, 2026
- License
- Apache-2.0
What is Onde Inference CLI MCP?
Onde Inference CLI is an MCP server. Fine-tune Qwen models with LoRA, export to GGUF, and manage local LLM deployments through a terminal interface or MCP server.
Summary generated from the project's README
About Onde Inference CLI MCP server
Onde Inference CLI provides a command-line interface for managing Onde Inference accounts, applications, and model catalogs. It allows developers to perform local fine-tuning of Qwen models using LoRA and export the resulting models to GGUF format. The tool is designed to work either as a terminal user interface or as an MCP server that communicates over stdio for use with AI agents.
Users can install the software through several package managers including Homebrew, npm, Cargo, pip, and NuGet. Once installed, the system supports managing local LLM deployments through the terminal. The CLI functions as a bridge between local hardware and the Onde Inference ecosystem, enabling specialized model training and management directly from a developer's workstation.
What you can do with it
- 1Fine-tune Qwen modelsA developer can instruct an AI agent to perform local fine-tuning on a Qwen model using LoRA.
- 2Manage Onde accountsAn agent can use the MCP server to handle account details and model catalogs.
- 3Operate via terminal UIUsers can interact with the system through a dedicated terminal interface for managing apps and models.
Written from the project's README. Every use case is checked against a quote from it.
How to install Onde Inference CLI MCP
Formats verified 10 Nov 2026Add to CursorAdd to GooseInstall Onde Inference CLI MCP in Claude Code
$ claude mcp add onde-cli -- npx -y @ondeinference/cliInstall Onde Inference CLI MCP in Cursor
~/.cursor/mcp.json
{
"mcpServers": {
"onde-cli": {
"command": "npx",
"args": [
"-y",
"@ondeinference/cli"
]
}
}
}Install Onde Inference CLI MCP in VS Code
.vscode/mcp.json
{
"servers": {
"onde-cli": {
"type": "stdio",
"command": "npx",
"args": [
"-y",
"@ondeinference/cli"
]
}
}
}Install Onde Inference CLI MCP in VS Code CLI
$ code --add-mcp '{"name":"onde-cli","type":"stdio","command":"npx","args":["-y","@ondeinference/cli"]}'Install Onde Inference CLI MCP in Claude Desktop
claude_desktop_config.json
{
"mcpServers": {
"onde-cli": {
"command": "npx",
"args": [
"-y",
"@ondeinference/cli"
]
}
}
}Install Onde Inference CLI MCP in Codex
$ codex mcp add onde-cli -- npx -y @ondeinference/cliInstall Onde Inference CLI MCP in Codex config.toml
~/.codex/config.toml
[mcp_servers.onde-cli] command = "npx" args = ["-y", "@ondeinference/cli"]
Install Onde Inference CLI MCP in Gemini CLI
$ gemini mcp add -s user onde-cli npx -- -y @ondeinference/cliInstall Onde Inference CLI MCP in Gemini settings.json
~/.gemini/settings.json
{
"mcpServers": {
"onde-cli": {
"command": "npx",
"args": [
"-y",
"@ondeinference/cli"
]
}
}
}Automated checks
Ran Nov 14, 2026Full report →Passed our automated checks. That is not a security audit: read the code and give the server the narrowest access it needs. How checks work
- Source repository linked and reachableondeinference/onde-cli is public and reachablegithub.com/ondeinference/onde-cli ↗Pass
- Package names this repository as its source@ondeinference/cli names ondeinference/onde-cli as sourcenpmjs.com/package/@ondeinference/cli ↗Pass
- Build provenance (Sigstore attestation)Sigstore attestation for @ondeinference/clinpmjs.com/package/@ondeinference/cli/v/0.5.1#provenance ↗Pass
- Open source license (OSI)Apache-2.0spdx.org/licenses/Apache-2.0.html ↗Pass
- Commit or release in the last yearLast commit or release 2026-11-10github.com/ondeinference/onde-cli ↗Pass
- OpenSSF ScorecardNot checked
- Package name not close to a popular packageNo popular package one edit away from @ondeinference/cliPass
- No critical advisory in the package or its dependencies7 resolved dependencies checked; no critical advisoryosv.dev ↗Pass
- No hidden instructions to the model0 tool descriptions and the README read; no override phrases, credential paths or hidden charactersPass
- Destructive tools are labelledNot checked
- Remote endpoint answers the MCP handshakeNot checked
README
View on GitHubonde is the command-line client for Onde Inference. It manages your
account, apps and model catalog, and it fine-tunes Qwen models locally with LoRA and exports them to
GGUF. It runs as a terminal UI, or as an MCP server for agents.
Install
Pick one:
# Homebrew
brew tap ondeinference/homebrew-tap && brew trust --tap ondeinference/homebrew-tap && brew install onde
# npm
npm install -g @ondeinference/cli
# Cargo
cargo install onde-cli
More ways, including pip/uv, .NET, Dart pub, winget, Scoop and pre-built binaries, are in docs/install.md.
Use it
onde --tuiopens the terminal UI where you sign in and manage apps and models (docs/usage.md).onde --mcpruns an MCP server over stdio for any MCP client (docs/mcp.md).- Fine-tune a model with LoRA and export it to GGUF (docs/finetune.md).
Docs
Install · Usage · MCP server · MCP Registry · Fine-tuning
License
Dual-licensed under MIT and Apache 2.0.
Copyright
© 2026 Splitfire AB (Onde Inference).
Onde Inference CLI MCP questions
How do I add Onde Inference CLI MCP to Claude Code?
Run claude mcp add onde-cli -- npx -y @ondeinference/cli in your terminal. Add --scope user to make it available in every project.
How do I add Onde Inference CLI MCP to Cursor?
Open ~/.cursor/mcp.json and add the server under mcpServers, using the Cursor config shown in the install section above.
What can I do with Onde Inference CLI?
You can manage your account, apps, and model catalog, fine-tune Qwen models locally using LoRA, and export them to GGUF format.
How can I use this with an AI agent?
You can run the tool as an MCP server over stdio to connect it to any MCP-compatible client.
Does it have a graphical interface?
It features a terminal user interface (TUI) for signing in and managing apps and models.