evalmonkey vs Recall

evalmonkey is an open-source framework for testing how LLM agents degrade under adversarial conditions. You run your agent against 10 standard datasets (GSM8K, ARC, HellaSwag, etc.) pulled automatically from HuggingFace, then apply chaos profiles that introduce realistic failure modes: malformed JSON schemas, artificial latency spikes, 429 rate-limit errors, context-window overflow, and prompt injection payloads. The key output is a degradation delta — evalmonkey shows you exactly how much your agent's accuracy drops under each failure type versus clean inputs. A model that scores 78% on GSM8K normally but drops to 31% when it gets a 429 mid-chain tells you something crucial about its error-recovery behavior that standard benchmarks completely miss. It supports OpenAI, Anthropic (via Bedrock and direct), Azure, GCP, and any Ollama-hosted model. Corbell-AI published this with a clear thesis: agents break in production for infrastructure reasons, not model reasons — and no existing benchmark tests that. evalmonkey was created today (April 17, 2026) and is still at 3 stars, but the core idea is genuinely novel in the evals space.

Recall is a local-first multimodal semantic search tool that lets you find any file on your computer using natural language — images, PDFs, audio, video, and text — without any manual tagging, folder organization, or metadata. Ask "that invoice from the dentist last spring" or "photo of the whiteboard with the Q3 roadmap" and it surfaces the right file. Under the hood, Recall uses Google's Gemini Embedding 2 to generate semantic embeddings for all your files and stores them in ChromaDB, a local vector database that runs entirely on your machine. Nothing leaves your device. The Raycast extension adds a visual grid UI so you can search from anywhere on macOS without opening a terminal. First-run indexing can take 20-30 minutes for large libraries, but subsequent queries are near-instant. The project is MIT-licensed and built by a solo developer. It's a clear response to the frustration that Spotlight, Find, and Windows Search still rely heavily on filename and metadata matching even in 2026. As Gemini Embedding 2 is free within generous limits, the operating cost is essentially zero for personal use.