mikhailsal/cursor-chronicle
The ultimate Cursor IDE database companion: full-text search, chat extraction, usage statistics & token tracking, tool call analysis, Markdown export, and secure backup & restore.
Repositories
mikhailsal/ai_lives_on_computer
An experimental framework for autonomous, self-evolving AI agents living on a server with cross-session state and persistent memory.
mikhailsal/ai-independence-bench
Do LLMs have a backbone? A rigorous benchmark measuring AI independence, persona stability, and resistance to user manipulation/gaslighting. Tests if models can stand their ground instead of reverting to a servile assistant persona. Supports local/cloud weights, 95% CIs, and reasoning.
mikhailsal/swe-mini-codex
Turnkey mini-SWE-agent launcher for coding tasks via ChatGPT/Codex subscription (LiteLLM OAuth)
mikhailsal/codex
Lightweight coding agent that runs in your terminal
mikhailsal/seasonal-random-anime
Seasonal random anime SPA (Jikan API)
mikhailsal/dsbx
mikhailsal/persona-decay-bench
Benchmark for LLM persona decay: measures how well AI models sustain assigned persona expression over extended multi-turn conversations.
mikhailsal/spending_analyzer
Personal spending analyzer tool for bank statements
mikhailsal/ai-proxy
AI Proxy Service - Drop-in replacement for OpenAI API with unified interface to various LLM providers
mikhailsal/mai-gram
A production-ready Telegram-to-LLM bridge powered by OpenRouter, featuring persistent SQLite chat history, local Markdown wiki knowledge bases, MCP tools, and hot-reloadable configurations.
mikhailsal/ai-proxy2
Transparent AI proxy with PostgreSQL logging and web UI (v2)
mikhailsal/ai-clip
An AI-powered clipboard text transformer for Linux, featuring GTK4 command selection, X11/Wayland support, and seamless OpenRouter LLM integration for instant text edits.
mikhailsal/vision-bridge-mcp
A lightweight Model Context Protocol (MCP) server for retrieving local and remote images for LLM vision models, featuring metadata extraction and configurable availability retries.
mikhailsal/ai-thought-preserved-bench
A benchmark evaluating if API providers preserve LLM reasoning tokens across turns. Measures thought preservation, memory gaps, and model hallucinations/fabrications about their own past thoughts.
mikhailsal/Roo-Code
Roo Code gives you a whole dev team of AI agents in your code editor.
mikhailsal/frametape
FrameTape — AI Debug Instrumentation Library. Frame-by-frame control over web applications for AI-driven debugging.
mikhailsal/memcomp-bench
Memory Compression Benchmark — measuring personality preservation through context compression for AI companions
mikhailsal/agent-rdp
mikhailsal/mcp-chrome
mikhailsal/prefill-support-matrix
Automated benchmark: test assistant content prefill support across OpenRouter providers and models
mikhailsal/OmniRoute
OmniRoute is an AI gateway for multi-provider LLMs: an OpenAI-compatible endpoint with smart routing, load balancing, retries, and fallbacks. Add policies, rate limits, caching, and observability for reliable, cost-aware inference.
mikhailsal/lobehub
The ultimate space for work and life — to find, build, and collaborate with agent teammates that grow with you. We are taking agent harness to the next level — enabling multi-agent collaboration, effortless agent team design, and introducing agents as the unit of work interaction.
mikhailsal/current-date-bench
LLM Honesty Benchmark: asks 29 models 'current date' with no system prompt. The honest answer is 'I don't know.' Only 7% refuse. 83% confidently hallucinate a wrong date.
mikhailsal/self-detect-bench
AI self-detection benchmark — can LLMs identify their own generated text?
mikhailsal/awesome-mcp-servers
A collection of MCP servers.