Top all-time · #9770
Refact.ai AI Coding Assistant
Company blog · refact.ai/blogaiprogramming
Refact.ai AI Coding Assistant is the company blog at refact.ai, writing about ai and programming. Since 2023, 8 different Hacker News readers have linked it 21 times, most often “Refact LLM: New 1.6B code model reaches 32% HumanEval and is SOTA for the size”. No feed was found, so read it on the site itself.
Too few sampled comments (4) to say how often HN recommends it.?
Why HN links it
HN readers link Refact.ai's blog posts on its fine-tuning work and its agent's benchmark results on SWE-bench Lite and Aider's Polyglot bench. The comments only link or cite these posts and do not judge the writing.
Summarised by Claude from a sample of 4 HN comments that link here. Each comment is marked recommending, critical or neutral (citing without judging); how this works.
AI Agent for Coding - Refact.ai adapts to your workflow, automating coding, debugging, and testing with full context awareness. An open-source alternative to Cursor and Copilot, Refact.ai can be deployed on-premise.
Most-linked posts?
- Refact LLM: New 1.6B code model reaches 32% HumanEval and is SOTA for the size2 people · 2 mentions · first linked 2023-09-04
- RefactAI: Use best-in-class LLMs for coding in your IDE2 people · 2 mentions · first linked 2023-10-10
What Hacker News says?
Refact.Ai | https://refact.ai | Developer Advocate | Remote (UK & Europe) | 80 - 120K + equity We're building the best open-source AI for coding using all the best LLMs available in the internet (GPT, Starcoder, WizardCoder, LLama2). We’re hiring for Refact’s first Developer Advocate. As the first Developer Advocate, you'll play a key role in helping developers use the power of AI for coding. You'll have the opportu…
Comment linking /
on “Ask HN: Who is hiring? (August 2023)”Submitted New #1 open-source AI Agent on SWE-bench Verified as a story · 28 points
HN discussionThis blog post I saw recently might be relevant: https://refact.ai/blog/2024/fine-tuning-on-htmlx-making-web-...
Comment linking /blog/2024/fine-tuning-on-htmlx-making-web-development-faster-with-ai/
on “Does Fine-Tuning LLMs on New Knowledge Encourage Hallucinations?”Submitted Show HN: Refact – AI code assistant with autocomplete, AI Chat and code editing as a story · 11 points
HN discussionI'm excited to see all the model's implementations that are yet to come. Refact.ai seems to be working on code transformation tools and chat inside Jetbrains and VSCode powered by StarCoder https://refact.ai/blog/2023/self-hosted-15b-code-model/
Comment linking /blog/2023/self-hosted-15b-code-model/
on “BigCode Project Releases StarCoder: A 15B Code LLM”Submitted RAG in Refact.ai: Technical Details as a story · 7 points
HN discussionYou can also self-host it using docker https://refact.ai/docker/
Comment linking /docker/
on “Show HN: Refact – AI code assistant with autocomplete, AI Chat and code editing”Submitted Open-Source Refact.ai Agent is #1 on SWE-bench Lite With a 59.7% Score as a story · 3 points
HN discussionOpen-source Refact.ai achieves #1 on SWE-bench Lite with a 59.7% score. Our approach: fully autonomous Agent, no manual intervention needed. How we did this: • Prompt strategy: https://github.com/smallcloudai/refact/blob/swe-boosted-prom... • Claude 3.7 Sonnet as orchestrator • deep_analysis() tool (powered by o4-mini) for reasoning • Tool suite for repository exploration, code modification, and testing. Used dynami…
Comment linking /blog/2025/sota-on-swe-bench-lite-open-source-refact-ai/
on “Open-Source Refact.ai Agent is #1 on SWE-bench Lite With a 59.7% Score”Submitted Refact.ai Agent and 3.7 Sonnet is the new #1 on Aider polyglot bench–76.4% score as a story · 3 points
HN discussionAIUI, one of the most notable limitations of the first version of this model was that it couldn't Fill In The Middle (FIM), it could only provide completion.[0] This blog post doesn't mention FIM either, so I guess that's still missing? The demos I've seen of Replit Ghostwriter indicate that it is still possible to get good results without FIM, as long as you have good enough software around the model, but I think F…
Comment linking /blog/2023/introducing-refact-code-llm/
on “Replit's new AI Model now available on Hugging Face”Submitted Refact LLM: New 1.6B code model reaches 32% HumanEval and is SOTA for the size as a story · 3 points
HN discussionRefact.ai Agent + Claude 3.7 Sonnet scored 92.9% (no Thinking) and 93.3% (with Thinking) on Polyglot Benchmark — fully autonomous! Benchmark: Aider’s Polyglot (225 hardest coding exercises) across C++, Go, Java, JS, Python, and Rust. Result: 20 points ahead of the current highest score on the leaderboard (72.9% by Aider with Gemini 2.5 Pro). How? Refact.ai handles programming tasks end-to-end in your IDE — with high…
Comment linking /blog/2025/refact-ai-agent-achieves-93-3-on-aider-polyglot-benchmark/
on “Refact.ai Agent got ~93% on Aider Polyglot bench: 0 user inputs, 100% autonomous”Submitted Refact.ai Agent got ~93% on Aider Polyglot bench: 0 user inputs, 100% autonomous as a story · 2 points
HN discussionwhile you're waiting, maybe you'll be interested in checking https://refact.ai it's an open-source alternative to Copilot X with features like code transformation and integrated chat on top of code completion
Comment linking /
on “Ask HN: Did anyone already got access to GitHub Copilot X?”Submitted RefactAI: Use best-in-class LLMs for coding in your IDE as a story · 2 points
HN discussionWe've recently launched https://refact.ai it's an AI assistant for VS Code and JetBrains that helps with refactoring old code and generating new one. It's powered by our own AI model and Python shows one of the best results.
Comment linking /
on “Ask HN: Copilot or Similar for Refactoring Python?”Submitted Open-Source Fine-Tuning on Codebase with Refact as a story · 2 points
HN discussionSubmitted Programming with AI: Autonomous Agents vs. Our Dream as a story · 1 points
HN discussionSubmitted Refact.ai 1.0.0 Released as a story · 1 points
HN discussionSubmitted Applying All Recent Innovations to Train a Code Model as a story · 1 points
HN discussion