Skip to main content
Glama
Sign In
deenesjakoruzh
Kobold MCP Server

Kobold MCP Server

by PhialsBasement
OverviewSchemaSourceRelated ServersDiscussionsAdmin
JavaScript
Remote

Kobold MCP Server

A Model Context Protocol (MCP) server implementation for interfacing with KoboldAI. This server enables integration between KoboldAI's text generation capabilities and MCP-compatible applications.

Features

  • Text generation with KoboldAI

  • Chat completion with persistent memory

  • OpenAI-compatible API endpoints

  • Stable Diffusion integration

  • Built on the official MCP SDK

  • TypeScript implementation

Related MCP server: nanobanana-mcp

Installation

npm install kobold-mcp-server

Prerequisites

  • Node.js (v16 or higher)

  • npm or yarn package manager

  • Running KoboldAI instance

Usage

import { KoboldMCPServer } from 'kobold-mcp-server';

// Initialize the server
const server = new KoboldMCPServer();

// Start the server
server.start();

Configuration

The server can be configured through environment variables or a configuration object:

const config = {
  apiUrl: 'http://localhost:5001' // KoboldAI API endpoint
};
const server = new KoboldMCPServer(config);

Supported APIs

  • Core KoboldAI API (text generation, model info)

  • Chat completion with conversation memory

  • Text completion (OpenAI-compatible)

  • Stable Diffusion integration (txt2img, img2img)

  • Audio transcription and text-to-speech

  • Web search capabilities

Development

  1. Clone the repository:

git clone https://github.com/yourusername/kobold-mcp-server.git
cd kobold-mcp-server

  1. Install dependencies:

npm install

  1. Build the project:

npm run build

Dependencies

  • @modelcontextprotocol/sdk: ^1.0.1

  • node-fetch: ^2.6.1

  • zod: ^3.20.0

  • zod-to-json-schema: ^3.23.5

Contributing

Contributions welcome! Please feel free to submit a Pull Request.

License

MIT License - see LICENSE file for details.

Support

For issues and feature requests, please use the GitHub issue tracker.

Available Tools

20 tools
kobold_abortB

Abort the currently ongoing generation

Inputs: apiUrl.

kobold_chatC

Chat completion (OpenAI-compatible)

Inputs: messages (required), apiUrl, temperature, top_p, +2 more.

kobold_completeC

Text completion (OpenAI-compatible)

Inputs: prompt (required), apiUrl, max_tokens, temperature, +2 more.

kobold_detokenizeC

Convert token IDs to text

Inputs: tokens (required), apiUrl.

kobold_generateC

Generate text with KoboldAI

Inputs: prompt (required), apiUrl, max_length, max_context_length, +6 more.

kobold_img2imgC

Transform existing image using prompt

Inputs: prompt (required), init_images (required), apiUrl, negative_prompt, +7 more.

kobold_interrogateC

Generate caption for image

Inputs: image (required), apiUrl.

kobold_last_logprobsB

Get token logprobs from the last request

Inputs: apiUrl.

kobold_max_context_lengthB

Get current max context length setting

Inputs: apiUrl.

kobold_max_lengthB

Get current max length setting

Inputs: apiUrl.

kobold_model_infoB

Get current model information

Inputs: apiUrl.

kobold_perf_infoC

Get performance information

Inputs: apiUrl.

kobold_sd_modelsB

List available Stable Diffusion models

Inputs: apiUrl.

kobold_sd_samplersB

List available Stable Diffusion samplers

Inputs: apiUrl.

kobold_token_countC

Count tokens in text

Inputs: text (required), apiUrl.

kobold_transcribeC

Transcribe audio using Whisper

Inputs: audio (required), apiUrl, language.

kobold_ttsC

Generate text-to-speech audio

Inputs: text (required), apiUrl, voice, speed.

kobold_txt2imgC

Generate image from text prompt

Inputs: prompt (required), apiUrl, negative_prompt, width, +5 more.

kobold_versionB

Get KoboldAI version information

Inputs: apiUrl.

kobold_web_searchC

Search the web via DuckDuckGo

Inputs: query (required), apiUrl.

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

  1. 20 tool updates
    • First observedkobold_abort
    • First observedkobold_chat
    • First observedkobold_complete
    • First observedkobold_detokenize
    • First observedkobold_generate
    • First observedkobold_img2img
    • First observedkobold_interrogate
    • First observedkobold_last_logprobs
    • First observedkobold_max_context_length
    • First observedkobold_max_length
    • First observedkobold_model_info
    • First observedkobold_perf_info
    • First observedkobold_sd_models
    • First observedkobold_sd_samplers
    • First observedkobold_token_count
    • First observedkobold_transcribe
    • First observedkobold_tts
    • First observedkobold_txt2img
    • First observedkobold_version
    • First observedkobold_web_search
Deploy Server
One endpoint for every MCP toolRayrun runs every MCP server behind one URL and holds the credentials. Each call is decided per client, tool, and argument, then recorded.ray.runAd

TDQS

C2.9/5.0

Scored across 20 tools

Disambiguation3/5

Most tools have distinct purposes (e.g., text generation vs. image generation vs. utility functions), but there is some overlap between kobold_chat and kobold_complete (both for text generation) and kobold_token_count and kobold_detokenize (both token-related). Descriptions help clarify, but an agent might occasionally confuse similar tools.

Naming Consistency5/5

All tool names follow a consistent snake_case pattern with a 'kobold_' prefix and descriptive suffixes (e.g., kobold_generate, kobold_model_info). This uniformity makes the set predictable and easy to navigate, with no deviations in style.

Tool Count3/5

With 20 tools, the count is borderline high for a single server, as it covers diverse domains like text generation, image processing, audio, and utilities. While each tool serves a purpose, the scope feels broad, potentially overwhelming for agents focused on specific tasks.

Completeness4/5

The tool set covers core AI functionalities well, including text and image generation, audio processing, and model management. Minor gaps exist, such as no explicit tool for updating settings or managing generation parameters beyond aborting, but agents can work around these with the available tools.

Maintenance

ActivityNo data
ResponsivenessOutdated data

Resources

Related MCP Connectors

Related MCP Servers

  • Outsource MCP

    gwbischof
    A
    license
    A
    quality
    D
    maintenance
    An MCP server that enables AI applications to access 20+ model providers (including OpenAI, Anthropic, Google) through a unified interface for text and image generation.
    2
    30
    MIT

  • mcp-server-gpt-chat

    codeChap
    F
    license
    Not graded
    quality
    C
    maintenance
    MCP server for OpenAI ChatGPT API. Enables chat completions, vision analysis, embeddings, and model listing through natural language.
    -

  • NovelAI Image MCP

    xinvxueyuan
    A
    license
    Not graded
    quality
    B
    maintenance
    An MCP server that exposes NovelAI image generation as tools for AI agents, supporting txt2img, img2img, inpaint, upscaling, Director tools, ControlNet annotation, tag suggestions, vibe encoding, and subscription queries via stdio or streamable-http transports.
    12
    MIT

Appeared in Searches