multimodal

Omni is a local-first semantic search app for macOS that indexes text, code, PDFs, images, audio, and video on-device. It supports multilingual search, private offline use, and exposes a local endpoint for agents to query indexed files.

Self-hosted LLM proxy that redacts personal data before requests reach model providers, including inside tool-call arguments and multimodal content. It supports OpenAI-compatible clients, runs locally with zero telemetry, and uses Presidio plus an optional privacy filter for broader PII and secret detection.

Google’s Gemma 4 12B is a multimodal model for local and developer-focused AI use, with native audio and vision support, agentic workflows, and broad tooling support.

An open 30B agentic model from Meta designed to run locally on consumer hardware for coding, function calling, and other always-on agent workflows.

Google’s Gemini 3.5 Flash announcement describes a frontier model aimed at agents and coding, with notes on speed, multimodal generation, and availability through the Gemini app, Google Antigravity, Gemini API, AI Studio, Android Studio, and Gemini Enterprise.

An OpenAI-compatible AI router that sends requests across 12+ providers and 100+ models with automatic fallback and multimodal routing.

An autonomous terminal coding agent built for Android Termux plus Windows, macOS, and Linux. The repo describes PLAN and AGENT modes, image input, a live web server, rollback, diagnostics, memory, and support for multiple AI providers.

Open-source multi-modal Claude Code client that runs locally and lets Claude compose across plugins, GUIs, documents, charts, wikis, automations, and messaging bridges.

Google’s Gemini API guide for the latest models, with model comparisons, pricing, quickstarts, and migration notes for Gemini 3.6 Flash and 3.5 Flash-Lite.

A vision-language foundation model for autonomous driving that combines 3D perception, driving VQA, and motion planning in one framework. The repo shows performance tables, model structure, and links to the paper, blog, and model downloads.

Google’s announcement for new Gemini models focused on agentic workflows, coding, multimodal work, and cybersecurity use cases. The page includes performance claims, pricing, safety notes, and example applications.

Federaide is a multimodal multi-agent orchestration harness for Mac, PC, and Android, with peer-to-peer agent routing, terminal-based code execution, voice input/output, Telegram access, scheduled routines, memory, and desktop control.

A lightweight CLI for running multiple local MLX models on Apple silicon. It serves OpenAI-compatible endpoints, supports both text-only and multimodal models, and can browse/download from Hugging Face.

KINETK is an AI infrastructure product that turns social-web content into a real-time IP graph for creators, communities, and narratives. The site highlights an API and MCP for agents, multimodal embeddings, and a query layer for grounded retrieval.

A detailed walkthrough for running a local coding agent on macOS with llama.cpp, Gemma 4, MTP speculative decoding, image support, and Pi as the agent interface.

OneRingAI is an open-source TypeScript library for building AI agents with integrations and graph memory. The repository shows a substantial codebase with docs, setup files, testing, MCP-related material, and references to supported models and multimodal capabilities.

Google Research paper describing a deployed multimodal defense system for detecting coordinated synthetic spam and abusive media at scale, using account relatedness, synthetic content classification, and LoRA-tuned LLMs.