SuperAPI AI Directory
Partner with us中文AI newsFavorites

All AI tools

35,738 results · Page 795 of 993

Open Lovable

AI website builders

Open Lovable is a free, open-source AI tool built by Mendable AI that can quickly convert any website into a React/Next.js application. It uses Firecrawl to scrape the target website's HTML, CSS, and layout, then leverages large language models such as Claude, GPT, and Groq to analyze the structure, generating clean TypeScript code with Tailwind CSS styling. It is released under the MIT License and supports self-hosting and full customization. Users need to configure the E2B Sandbox, Firecrawl, and at least one AI model API key themselves, and can use an interactive AI chat interface to adjust styles and layout in real time using natural language.

Read more

Open Voice OS

AI audio

Open Voice OS (OVOS) is a community-driven open-source voice AI platform for creating custom voice-controlled interfaces on a variety of devices. It provides natural language processing (NLP), customizable user interfaces, and focuses on privacy and security. The project is maintained jointly by developers from the global Linux and open-source community, with the aim of making cutting-edge voice assistant technology available to everyone. It supports embedded headless devices and single-board computers with screens, can be used to build DIY smart speakers, and can also serve as an experimental platform for new features. OVOS is completely free and open source, released under the Apache 2.0 license, allowing personal or commercial use, modification, and distribution. By default, ovos-core does not require a backend internet server to run, but the default speech-to-text engine currently requires an internet connection, while local offline options such as Vosk and Mozilla DeepSpeech are also available. Text-to-speech supports Mimic3 (local), Google TTS, Amazon Polly, and others. Installation methods include Docker and Python, and prebuilt images are also provided for Raspberry Pi and Mycroft® devices.

Read more

Open-O1

AI developer platforms

Open-O1 is an open-source project aimed at reproducing the long reasoning and problem-solving capabilities of the OpenAI O1 model through open-source innovation. The project compiles a set of SFT data for chain-of-thought (CoT) activation and uses it to train LLaMA and Qwen models, enabling smaller models to achieve enhanced long reasoning performance. The project provides two models, OpenO1-LLaMA-8B-v0.1 and OpenO1-Qwen-7B-v0.1, as well as the OpenO1 SFT data for CoT activation, all available on HuggingFace. Official evaluations show that its models achieve competitive results on benchmarks such as GSM8K, MATH, MMLU, ARC-C, and BBH, with some metrics outperforming llama3.1-8b-instruct. The project is non-profit, welcomes community participation, and will continue to open-source the code and materials from the development process.

Read more

open-r1-reprod

AI developer platforms

open-r1-reprod is a reproduction project of open-r1, using GRPO to train Qwen models of 0.5B, 1.5B, 3B, and 7B on the MATH-lighteval math dataset, observing and analyzing phenomena such as the reasoning abilities, rule reward hacks, and multilingual mixed reasoning of models of different sizes. The project code is based on open-r1, modifying the config_grpo.yaml configuration file, training uses a cosine learning rate (maximum 3.0e-06), and the rewards include accuracy reward, format reward, and tag_count_reward. The project is not exactly a success, but there are some interesting findings, for example, the 0.5B small model is very difficult to learn reasoning ability through GRPO, models of 1.5B and above begin to take effect; the 7B model's loss curve fluctuates violently, and phenomena such as garbled output appear.

Read more

Open-R1-Video

AI video

Open-R1-Video is the first open-source R1-like video large language model project, aiming to introduce the R1 reinforcement learning paradigm into video understanding tasks. The project is based on the Qwen2-VL-7B-Instruct model and uses GRPO (pure reinforcement learning, without annotated reasoning traces) for training on 4 A100 (80G) GPUs, with training data consisting only of videos, questions, and the correct answer option letters. The project has open-sourced the training code, data (open-r1-video-4k), as well as inference scripts and evaluation results, and has released the temporary model Open-R1-Video-7B. This model improves video understanding capabilities through forced-thinking GRPO training, and the training process is recorded in Wandb logs for reference. The project also provides a data reformatting tool, a method for evaluation using Lmms-eval, and supports single-GPU A100 training.

Read more

Open-Reasoner-Zero

AI learning

Open-Reasoner-Zero is the first open-source implementation of large-scale reasoning-oriented reinforcement learning training, emphasizing scalability, simplicity, and accessibility. It uses the same base model as DeepSeek-R1-Zero-Qwen-32B and achieves better performance on the AIME2024, MATH500, and GPQA Diamond benchmarks while requiring only one-tenth of its training steps. The project has open-sourced the source code, parameter settings, training data, and model weights, provides multiple model sizes from 0.5B to 32B, and supports training the 0.5B model on a single A800/H800 GPU, making it easy for researchers and developers to participate in reasoning model training.

Read more

open-seo

AI coding

OpenSEO is an open-source SEO tool positioned as an alternative to Semrush and Ahrefs, using a pay-as-you-go model controlled by users themselves. It provides focused workflows such as keyword research and competitor insights, and comes with a built-in MCP server and Agent Skills that can connect to AI agents such as Claude Code, OpenClaw, and Hermes. Users bring their own DataForSEO API Key and pay DataForSEO directly to use SEO data. It supports two self-hosting approaches, Docker and Cloudflare, and also offers a hosted subscription service for $10 per month, with a free trial available on the official website.

Read more