👋 Welcome to Forever AI

Build without limits.

We are Forever AI (formerly Klox AI) — an open-source AI lab building large language models, image generation tools, and hands-on Spaces that anyone can use, study, and improve.

We believe great AI shouldn't live behind closed doors. So we build in the open, ship real tools, and share everything we learn with the community.

---

🧠 About Us

Forever AI is a small, focused team of engineers and researchers who care about useful, reproducible, community-friendly AI.

We don't chase hype — we ship models, datasets, and demos that solve real problems. Everything we publish is documented, benchmarked, and designed to be picked up and used by other developers.

Our work spans the full stack: from training loops and dataset pipelines to inference servers and interactive Spaces.

---

🚀 What We Do

🦍 Large Language Models

We train and release our own LLMs under the Ape's line — starting with Apeiron 1.0, a coding-focused agent trained from scratch. Each release is a measurable step in loss, accuracy, and task coverage.

• From-scratch training, no borrowed base weights
• Coding, reasoning, and agentic workflows
• Full model cards with honest limitations

🎨 Image Generation

We experiment with diffusion models, fine-tunes, and ControlNet-style pipelines — with an emphasis on style control, prompt fidelity, and reproducibility.

• Custom LoRAs and fine-tunes
• Prompt engineering guides
• Ready-to-use Gradio demos

🧪 Spaces & Tools

We build interactive Python Spaces on Hugging Face so anyone can try our models in the browser — no setup, no GPU, no excuses.

• Gradio and Streamlit demos
• API-ready inference endpoints
• Open notebooks and example code

📚 Open Datasets

We publish the datasets we train on whenever licensing allows — because good models deserve transparent data.

---

🌍 Our Mission

Make artificial intelligence accessible to everyone.

That means:

• 🔓 Open by default — weights, code, and docs whenever possible
• 📖 Honest about limits — no inflated benchmarks, no vague claims
• 🧰 Useful over flashy — tools people actually reach for
• 🤝 Community-first — feedback shapes what we build next

---

📬 Contact & Links

---

🤝 Get Involved

We love open-source contributors. Here's how you can help:

• ⭐ Star our repos and models you find useful
• 🐛 Report bugs and open issues
• 🧪 Try our Spaces and share feedback
• 🔧 Submit PRs — docs, fixes, and features all welcome
• 📣 Spread the word — tell a friend who builds

---

📜 License & Ethics

Unless stated otherwise, our models and code are released under open licenses (Apache 2.0, MIT, or similar). We ask that you:

• Respect the license of each individual release
• Verify generated code and outputs before production use
• Follow the laws and regulations of your jurisdiction
• Use our tools to build things that help people

---

Built with care by Forever AI — for everyone. 🦍