Got nerfed? Here is the free version.
Pick your nerfed tool. We list the open-source replacement that does the same job for $0 (or close to it). Manually curated, no affiliate links, no recommendations driven by anything except actual fit.
Star counts are approximate, refreshed weekly via the GitHub API. List last human-reviewed 2026-06-04. Suggest an addition or correction: open a PR.
Default Midjourney replacement if you have a GPU. Largest community + extension library.
git clone, run webui-user.bat / webui.sh- Ollama ↗130k
Easiest way to run a local LLM. Pair with Open WebUI for full ChatGPT replacement.
Run Llama/Mistral/Gemma/DeepSeek locallyOllama-compatible API for any clientBrew/winget/curl installINSTALLbrew install ollama && ollama run llama3.3CAVEATNeeds 16GB+ RAM for usable models. GPU strongly recommended.Open repo → - OpenCode ↗120k
Open-source Claude Code / Cursor replacement that lives in the terminal. Pick when you want zero vendor lock-in and any LLM backend.
Terminal-native TUI coding agent (vim-like)Build/Plan mode toggle, session managementProvider-agnostic: Claude/OpenAI/Google/localGitHub app for issues + PRsINSTALLcurl -fsSL https://opencode.ai/install | bash (or npm i -g opencode-ai@latest)CAVEATBYO API key (or local model). No GUI editor - it is a terminal TUI.Open repo → - Open WebUI ↗95k
The default ChatGPT-replacement self-host. Massive ecosystem.
ChatGPT-style web UI for any backend (Ollama, OpenAI-compatible)RAG with file uploadsFunction calling, web search pluginINSTALLdocker run -p 3000:8080 ghcr.io/open-webui/open-webui:mainCAVEATBest paired with Ollama running locally.Open repo → - GLM-5.2 (open weights) ↗
Permissive (MIT) open-weight model targeting ChatGPT/Claude-class chat and agentic coding without API lock-in - the strongest open option for a self-hosted stack.
Fully-open (MIT) flagship MoE from Z.ai / Zhipu AI (weights released Jun 2026)~744B total / ~40B active parameters, 1M-token context (~5x GLM-5.1)Strongest open-weight coding model at release - #2 on Code Arena, ~#3 on FrontierSWE behind only Fable 5 / Opus 4.8Tool use + reasoning, served via vLLM/SGLangINSTALLollama run glm-5.2 (or download weights from HuggingFace: zai-org/GLM-5.2, serve via vLLM/SGLang)CAVEATMIT, fully open, no usage restrictions. At ~744B params it is multi-GPU / data-center scale - use a hosted provider if you lack the hardware. Note: the hosted Z.ai API routes through China (a data-residency consideration); self-hosting the open weights avoids that.Open repo → - DeepSeek-V4 (open weights) ↗
Drop-in open-weight replacement for the OpenAI/Anthropic chat APIs. Flash is the locally-runnable variant; serve it behind Open WebUI.
MIT-licensed open-weight MoE: V4-Flash (284B/13B active) + V4-Pro (1.6T/49B active)1M-token contextNear-frontier quality at a fraction of API costRun locally or self-host the backendINSTALLDownload from HuggingFace (deepseek-ai/DeepSeek-V4-Flash); serve via vLLM/SGLangCAVEATEven Flash (284B) needs serious VRAM/multi-GPU; Pro (1.6T) is data-center scale. MIT license, weights fully open.Open repo → - ComfyUI ↗65k
When you want full control over the diffusion pipeline. Power-user choice.
Node-based image/video pipelineSDXL, Flux, video diffusion modelsReproducible workflows you can share as JSONINSTALLgit clone + Python depsCAVEATSteep learning curve vs A1111.Open repo → - AppFlowy AI ↗65k
Notion + Notion AI in one self-hostable package.
Open-source Notion alternativeAI features integratedSelf-host or desktop appINSTALLDownload desktop app from appflowy.io or self-host serverCAVEATAI features require either AppFlowy Cloud subscription or BYO OpenAI key.Open repo → - Mistral Large 3 (open weights) ↗
Unrestricted (Apache-2.0) open-weight model to replace the OpenAI/Anthropic/Gemini chat APIs in a self-hosted stack.
Apache-2.0 open-weight MoE from Mistral AI: 675B total / 41B active256K context, multimodal (text + vision)Production-grade general-purpose modelServe via vLLM/SGLang or run through OllamaINSTALLollama run mistral-large-3 (or download weights from HuggingFace: mistralai/Mistral-Large-3-675B-Instruct-2512-BF16, serve via vLLM/SGLang)CAVEATApache-2.0, no usage restrictions. December 2025 release. At 675B params it is multi-GPU / data-center scale - use a hosted provider if you lack the hardware.Open repo → - Open Interpreter ↗60k
ChatGPT's Code Interpreter, but local and against any model.
Code Interpreter for any LLM, locallyFile ops + shell + browser controlPython-first but multi-languageINSTALLpip install open-interpreterCAVEATAGPL - be careful if integrating into proprietary product.Open repo → - Qwen3.6 (open weights) ↗
Permissive (Apache-2.0) open-weight model to back a self-hosted ChatGPT/Claude replacement. The 35B-A3B MoE runs on a single high-RAM machine.
Apache-2.0 open-weight family from Alibaba's Qwen teamQwen3.6-35B-A3B MoE (35B total, 3B active) + 27B dense256K context, extensible to ~1MStrong agentic + coding scoresINSTALLollama run qwen3.6 (or pull from HuggingFace: Qwen/Qwen3.6-35B-A3B)CAVEATLarger variants need a GPU; the 35B MoE wants ~24GB+ to be comfortable.Open repo → - Google Gemma 4 (open weights) ↗
Local, commercially-usable alternative to the Gemini/GPT chat APIs. The 2B/4B run on a laptop; 31B is the flagship.
Apache-2.0 open-weight multimodal models from Google DeepMindVariants 2B/4B/26B-MoE/31B-denseNative audio + vision, function callingRuns offline on phones, Jetson, consumer GPUsINSTALLollama run gemma4 (variants e2b/e4b/12b/26b/31b)CAVEATApache-2.0 (note: earlier Gemma releases used the custom Gemma license; Gemma 4 is Apache-2.0). Bigger variants need a GPU.Open repo → - LobeChat ↗55k
Slickest UI of the open-source chat clones.
Polished chat UI with plugin storeMulti-provider (OpenAI/Anthropic/Gemini/local)Vision, voice, knowledge basesINSTALLVercel one-click deploy or docker composeCAVEATBest with API-key access; some features paywall on hosted version.Open repo → - Kimi K2.6 (open weights) ↗
Near-frontier open-weight model to back a self-hosted ChatGPT/Claude replacement or an agentic coding stack.
Open-weight multimodal MoE from Moonshot AI: 1T total / 32B active256K context, handles text + image + videoINT4-native weights for cheaper servingStrong agentic + coding performanceINSTALLollama run kimi-k2.6 (or download weights from HuggingFace: moonshotai/Kimi-K2.6, serve via vLLM/SGLang)CAVEATModified MIT: commercial use is free, but if your product exceeds 100M monthly active users OR $20M/month revenue you must display "Kimi K2.6" in the product UI. At 1T params it is data-center scale - use a hosted provider (e.g. OpenRouter) if you lack the GPUs.Open repo → - NVIDIA Nemotron 3 Nano (open weights) ↗
Efficient open-weight model backend for a self-hosted chat/agent stack, as an alternative to paid frontier APIs.
Open-weight hybrid Mamba-Transformer MoE from NVIDIANano: ~31.6B total / 3.2B active1M-token context, built for agentic reasoningWeights, recipe + data released on HF/NGCINSTALLollama run nemotron3 (or HuggingFace: nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16)CAVEATNVIDIA Open Model License (permissive, commercial use allowed) - read it, it is not OSI Apache/MIT. GPU recommended.Open repo → - FLUX.2 [klein] 4B ↗
Fast, fully-open (Apache-2.0) Midjourney alternative you can run locally. The 4B 'klein' is the commercially-usable variant.
4B rectified-flow transformer from Black Forest LabsText-to-image + multi-reference editing in one modelSub-second generation on consumer GPUsLoads in ComfyUI or the official flux2 repoINSTALLDownload FLUX.2-klein-4B from HuggingFace; run in ComfyUI or via github.com/black-forest-labs/flux2CAVEATOnly the 4B klein is Apache-2.0 - the 9B klein and FLUX.2 [dev] are non-commercial. Needs ~13GB VRAM (RTX 3090/4070+).Open repo → - MiniMax M2.7 (open weights) ↗
A capable open-weight model for personal, self-hosted, or research chat/coding setups - if your use is non-commercial.
229B open-weight model from MiniMaxBuilt for coding + agentic workflows (tool calling, agent teams)Served via vLLM/SGLang/TransformersWeights published on HuggingFaceINSTALLollama run minimax-m2.7 (or download weights from HuggingFace: MiniMaxAI/MiniMax-M2.7, serve via vLLM/SGLang)CAVEATNOT OSI open-source. The MiniMax M2.7 Non-Commercial License permits personal, self-host, and research use only - any commercial use requires prior written authorization from MiniMax (api@minimax.io), and commercial users must display "Built with MiniMax M2.7". At 229B params, plan for multi-GPU serving or a hosted provider.Open repo → - Fooocus ↗42k
If you want Midjourney's simplicity without ComfyUI's complexity.
Midjourney-feel UX (one prompt, presets)Built-in style libraryMinimal config - just runINSTALLDownload zip, run run.bat / run.shCAVEATLess customizable than A1111/ComfyUI.Open repo → - OpenHands ↗38k
Open-source Devin. Best when you want autonomous multi-step coding agents you control.
Sandboxed agent that runs full dev tasksWeb UI + CLI + headless modesPlug any LLM backendINSTALLdocker run docker.all-hands.dev/all-hands-ai/openhands:0.30CAVEATDocker required. Heavier setup than CLI tools.Open repo → - Coqui TTS ↗38k
Production-grade TTS toolkit. Bigger than just one model.
Mature TTS toolkit (XTTS v2)Fine-tune on your own voice1100+ language support via modelsINSTALLpip install TTSCAVEATCoqui Inc. shut down - community fork active.Open repo → - OpenHands (also covers v0/Bolt) ↗38k
Closest open-source v0/Bolt clone. Sandboxed agent that builds apps.
Sandbox runs full coding tasks end-to-endBrowser + terminal + code-edit primitivesPlug any LLMINSTALLDockerCAVEATHeavier setup than browser-based v0/Bolt. BYO LLM API key.Open repo → - ChatTTS ↗35k
Best free ElevenLabs replacement for English/Mandarin chat-style audio.
Conversational TTS - handles fillers, pauses, multi-speakerMandarin + EnglishReal-time on consumer GPUINSTALLpip install ChatTTSCAVEATAGPL license - commercial use restricted.Open repo → - AnythingLLM ↗32k
When the use case is 'chat with my company docs' rather than general LLM.
Workspace-based RAG over your docsDesktop or self-host20+ vector DB integrationsINSTALLDesktop app or DockerCAVEATDesigned around document ingestion, not general chat. BYO LLM key.Open repo → - aider ↗28k
Best Cursor replacement for terminal-comfortable devs who want auditable git history of every edit.
CLI-first, terminal-nativeGit-aware: auto-commits, branches, diffsWorks with any LLM provider (BYO API key)Strong repo-map contextINSTALLpip install aider-chatCAVEATBYO API key (Anthropic/OpenAI/local). No GUI.Open repo → - Jan ↗28k
Easiest desktop ChatGPT replacement for non-technical users.
Desktop app - no Docker, no terminalBuilt-in model marketplaceLocal-first, optional cloudINSTALLDownload from jan.aiCAVEATAGPL. Mac/Win/Linux.Open repo → - LTX-2 (Lightricks) ↗
Open-weight Runway/Sora/Veo alternative that generates video AND matching audio in one model.
DiT-based foundation model with synchronized audio + videoText/image/audio-to-video, up to 4KOpen weights + training code + LoRA trainerRuns locally (FP8 quantization for lower VRAM)INSTALLgit clone https://github.com/Lightricks/LTX-2 && uv sync --frozenCAVEATCommunity License, NOT Apache/MIT: free (incl. commercial) only under $10M annual revenue; above that needs a paid commercial license. Heavy GPU for HD/4K.Open repo → - LibreChat ↗25k
When you want ChatGPT's UX but as a self-hosted multi-user product.
ChatGPT clone with multi-provider supportPlugins (DALL-E, Wolfram)Auth + multi-userConversation forkingINSTALLDocker composeCAVEATHeavier setup than Open WebUI.Open repo → - Cline ↗24k
Direct Cursor/Claude-Code clone inside VSCode. Best for devs already living in VSCode.
VSCode extension, native UXPlan/Act mode toggleBuilt-in MCP supportPermission-based file opsINSTALLVSCode marketplace: 'Cline'CAVEATBYO API key. Permission prompts can feel noisy.Open repo → - Tabby ↗24k
When you need 100% on-prem code completion. No data leaves your machine.
Self-hosted, runs on your GPUOpenAPI-compatibleTeam collab featuresINSTALLDocker image or HomebrewCAVEATRequires GPU for decent latency. Setup is non-trivial.Open repo → - InvokeAI ↗24k
Cleanest UI in the SD ecosystem. Best for design-pipeline use.
Polished web UICanvas/inpainting workflowStudio-grade install experienceINSTALLInstaller .exe / .pkg / launcher scriptCAVEATNeeds 8GB+ VRAM. Smaller extension ecosystem than A1111/ComfyUI.Open repo → - MOSS-TTS (OpenMOSS) ↗
Best free ElevenLabs alternative with a genuinely permissive license - Apache-2.0, so commercial use is allowed (unlike AGPL ChatTTS).
Apache-2.0 speech/sound family from OpenMOSS / MOSI.AIZero-shot voice cloning + multi-speaker dialogue (MOSS-TTSD)Long-form, multilingual, real-time streamingTiny Nano-100M variant runs on CPUINSTALLgit clone https://github.com/OpenMOSS/MOSS-TTS && pip install -e ".[torch-runtime]"CAVEATLarger variants want a GPU; the Nano-100M is the CPU-friendly option.Open repo → - Open-Sora ↗22k
When you want OpenAI Sora capability locally. Heavy GPU need.
Open-source video diffusion (Sora-style)Text-to-video and image-to-videoPublic model checkpointsINSTALLgit clone + pip install -r requirements.txtCAVEATNeeds 24GB+ VRAM realistically.Open repo → - AudioCraft (MusicGen) ↗22k
Best free Suno/Udio alternative for music generation.
Meta's music + audio generation toolkitText-to-music via MusicGenAudioGen for sound effectsINSTALLpip install -U audiocraftCAVEATOutput quality below Suno's; needs prompt engineering.Open repo → - Continue.dev ↗21k
Best when you want local LLMs (via Ollama) and care about privacy. Strong autocomplete.
VSCode + JetBrains supportLocal-model friendly (Ollama integration)Custom slash commandsTab autocomplete + chatINSTALLVSCode/JetBrains marketplace: 'Continue'CAVEATTab autocomplete needs a fast local model or hosted API.Open repo → - Perplexica ↗19k
Direct Perplexity replacement. Free if you BYO API key.
Open-source Perplexity cloneMultiple search modes (Academic, YouTube, Reddit)Self-host, BYO LLM + SearXNGINSTALLDocker compose with SearXNGCAVEATRequires running SearXNG alongside. Self-host setup is ~30 minutes.Open repo → - Stable Video Diffusion ↗
Best image-to-video model with permissive-enough license.
Image-to-video model from StabilityRun via ComfyUI workflowFree for non-commercial; commercial license availableINSTALLDownload checkpoint from HuggingFace, load in ComfyUICAVEATRead the license - not pure open-source.Open homepage → - SearXNG ↗17k
Use under Perplexica/Morphic for the search backend, or just as a private Google.
Privacy-respecting meta-search engine70+ search backendsFoundation for AI search toolsINSTALLDockerCAVEATDefault config tunes for privacy, not result quality. Tweak engines/weights for AI search backing.Open repo → - Bolt.diy ↗17k
Direct Bolt.new replacement. Free if you bring your own API key.
Open-source fork of Bolt.new (StackBlitz)WebContainers in browserBYO API key (any provider)INSTALLgit clone + npm installCAVEATWebContainer model limits some Node packages. BYO API key for any LLM provider.Open repo → - Goose ↗14k
When you want a polished agent from a real engineering org. MCP-native.
Block (Square)'s on-machine AI agentExtension marketplace (MCP servers)Fast, focused on dev workflowsINSTALLbrew install block/goose/gooseCAVEATNewer than aider/Cline; ecosystem still growing.Open repo → - F5-TTS ↗14k
Voice cloning use case. MIT-licensed (rare for high-quality TTS).
Voice cloning from a 10-second sampleMulti-languageStrong prosodyINSTALLpip install f5-ttsCAVEATBest results with high-quality 10s reference audio. Mandarin support stronger than other languages.Open repo → - Roo Code ↗11k
Cline's multi-personality cousin. Pick when you want explicit Architect-vs-Coder separation.
Cline fork with extra modes (Architect, Ask, Debug)Multi-mode workflowsCustom mode authoringINSTALLVSCode marketplace: 'Roo Code'CAVEATSame BYO-API-key model as Cline.Open repo → - Morphic ↗8.5k
Lighter-weight Perplexica with prettier output.
Generative UI search engineStreaming results with rich cardsVercel-deployable in 1 clickINSTALLgit clone + Vercel deploy or npm run devCAVEATSmaller community.Open repo → - Codeium ↗
Closest free-tier replacement for Copilot. No usage caps on free.
Free tier with no token cap70+ language supportMulti-IDE (VSCode/JetBrains/Vim/Emacs)Self-host enterprise tierINSTALLVSCode/JetBrains marketplace: 'Codeium'CAVEATClosed-source on the free tier. Enterprise plan exists for self-host.Open homepage →