PseudoLlama Server

A tool that forwards responses through an Ollama-like server. Allows you to pipe through responses from LM Studio or OpenRouter to any Ollama-compatible endpoint.

Overview

PseudoLlama is a simple Express server that mimics the Ollama API. It serves content from a text file as responses to API requests, making it useful for testing applications that integrate with Ollama.

Features

• Simulates Ollama API endpoints (/api/chat, /api/generate, etc.)
• Also supports OpenAI-compatible endpoints (/v1/chat/completions, /v1/completions, etc.)
• Web UI for editing the content and testing the server
• Supports both streaming and non-streaming responses
• Comprehensive logging of all model communications (requests and responses)

Installation

npm install

Usage

Start the server:

npm start

The server runs on port 12345 by default. This is a fixed port for testing purposes.

IMPORTANT: When connecting to this server from other tools, you must specify port 12345 in your configuration.

API Endpoints

Ollama-style Endpoints

• POST /api/chat - Chat completions
• POST /api/generate - Text generation
• POST /api/embeddings - Generate embeddings
• GET /api/tags - List available models
• POST /api/pull - Simulate model pulling

OpenAI-compatible Endpoints

• POST /v1/chat/completions - Chat completions
• POST /v1/completions - Text completions
• POST /v1/embeddings - Generate embeddings
• GET /v1/models - List available models

Server Management

• GET /api/server/status - Check server status
• POST /api/server/toggle - Enable/disable the server
• GET /api/content - Get the current content
• POST /api/content - Update the content

Web UI

Access the web UI by navigating to http://localhost:12345 in your browser. The UI allows you to:

• View and edit the content that will be returned by the API
• Test the API by sending a request to the server
• Enable/disable the server

Logging

PseudoLlama includes comprehensive logging of all model communications:

Console Logging

Basic request and response information is logged to the console when the server is running.

Full Communication Logging

Complete model communications (including full request and response bodies) are logged to logs/model_communications.log. This is particularly useful for:

• Debugging applications that integrate with language models
• Analyzing the exact data sent to and received from models
• Understanding the structure of streaming responses

Log Viewer Utility

A log viewer utility is included to help analyze the logs:

# View all logs
node view-logs.js

# Show only the last 10 log entries
node view-logs.js --limit=10

# Filter logs by model
node view-logs.js --model=openrouter

# Filter logs by endpoint
node view-logs.js --endpoint=/v1/chat

# Show only requests
node view-logs.js --requests

# Show only responses
node view-logs.js --responses

# Watch for new log entries in real-time
node view-logs.js --tail

# Show help
node view-logs.js --help

The log files are automatically rotated when they reach 10MB to prevent excessive disk usage.