Useful free-to-run and open-source projects, reviewed for license, activity, and practical requirements. A public repository does not guarantee free hosted inference, security, or suitability for every use.
Showing 1–9 of 13 verified projects
AI Image & Design
ComfyUIFeatured
Comfy-Org/ComfyUI
Build image and video generation workflows by connecting model and processing nodes.
A modular interface and inference backend for composing diffusion workflows. The README documents local installation on Windows, Linux, and macOS with support for multiple GPU types; Comfy Cloud is a separate paid hosted option. Hardware needs depend on the models and workflows selected.
Terminal coding agent with editor integrations and support for hosted or local model providers.
An open-source coding agent that can work from the terminal and related clients. The README documents Linux, macOS, and Windows installation and support for Qwen, OpenAI-compatible, Anthropic, Gemini, third-party, and local model providers such as Ollama or vLLM. API credentials and inference costs depend on the provider selected.
Desktop and terminal AI agent for coding and multi-step computer workflows.
A native desktop app, CLI, and API for AI-assisted development and task execution. The README lists multiple hosted providers, Ollama, and subscription-based provider connections. Goose is part of the Linux Foundation Agentic AI Foundation; provider credentials and costs vary by route.
Download and run open-weight language models locally or connect to Ollama Cloud.
Ollama packages model downloads and local inference behind a CLI and API. The README includes installers for macOS, Windows, and Linux, plus Docker guidance. Local model availability and performance depend on model size and machine resources; cloud is a separate service.
Offline desktop speech-to-text that inserts transcriptions into the active text field.
A cross-platform transcription application designed to run locally. The README describes Whisper model options with optional GPU acceleration and Parakeet V3 with CPU-optimized inference. It supports Windows, macOS, and Linux and does not require sending speech to a hosted transcription service.
Python and TypeScript tooling for controlling browsers with AI agents.
The repository contains open-source browser-agent libraries that can run with a local browser and a chosen LLM. The README also promotes separate hosted browsers, proxy, and agent services with usage-based pricing. Credentials and costs depend on the model and whether hosted infrastructure is selected.
Crawl, scrape, search, and convert web content into formats for AI applications.
A web-data extraction project with API clients and a hosted service. The README quick start uses a hosted Firecrawl API key and describes paid/limited hosted usage; the code can also be self-hosted under AGPL-3.0, with deployment and external-service requirements documented separately.
Python framework for composing retrieval, search, and LLM application pipelines.
A modular framework for RAG, semantic search, agents, and multimodal LLM applications. The README documents Python installation and configurable pipelines. The framework can be self-hosted; model and data connectors may require external credentials or paid services depending on configuration.
Orchestrate data, infrastructure, business, and AI workflows using declarative flows.
A workflow orchestration platform with event-driven execution, plugin integrations, and AI agent/RAG workflow documentation. The README documents self-managed infrastructure and separate cloud deployment paths. Workflows may call paid external APIs or services, but those are not required for the base orchestrator.
Independent community-built projects that may be useful to explore, but have unresolved license, security, or service-term questions. They are not endorsements and are excluded from verified counts, filters, and recommendations.
AI Image & Design
Open Higgsfield AI
Autom8AI/Open-Higgsfield-AI
Community · Unverified
Independent image and video studio UI that sends generation requests to MuAPI.
A community project with Higgsfield-inspired workflows. Its README identifies MuAPI as the inference provider, requires a MuAPI API key for local setup, and advertises a separate hosted MuAPI version. This is not an official Higgsfield product, and the repository README claims MIT while GitHub reports no detected license and the repository has no root LICENSE file.
Unverified community project. This is not an official Higgsfield product. The README claims MIT, but GitHub does not detect a license and no root LICENSE file was found. Local setup routes generation through MuAPI and requires its API key; the README's privacy and request-flow statements conflict. The README also describes its installers as unsigned/not notarized. Review the repository, license, security, and service terms before installing.
Also in: AI Video Generation
community projectMuAPIimage generationvideo generationthird-party
Listings are reviewed for public availability, current repository status, a clear license, recent maintenance, and a README that supports the stated use. We record hardware, credentials, and paid-service dependencies when documented. This is an editorial selection, not an endorsement, security audit, or guarantee of future maintenance. Verified projects appear in the main results. The separate community section is explicitly unverified and is not an endorsement.
Repository metadata is fetched from GitHub on request and cached server-side. Editorial categories, requirements, and verification notes are held separately; live metadata does not automatically promote or reclassify projects.