Start a real LLM training in 10 minutes. Tell Claude Code or Codex what the model is for; they prepare the datasets and create the training project. Click "Run Training" in the local studio. Arkor runs the training, and deploys the trained model as an OpenAI-compatible API. Think Next.js and Vercel for fine-tuning: code you can review, infrastructure you do not have to manage, and a model your app can call. No GPU setup. No ML expertise required. No Python training code to write.
We built Arkor because we kept running into the same gap: adding an AI feature to an application is easy, but adapting a model to a specific task still feels like joining an ML infrastructure team.
The moment you want to fine-tune, you often end up maintaining a separate Python project, converting datasets into unfamiliar formats, provisioning GPUs, and moving between scripts, notebooks, and dashboards.
We wanted model development to feel more like ordinary software development.
With Arkor, the workflow starts inside your existing repository:
Tell Claude Code or Codex what behavior you want from the model.
Your coding agent can find or prepare a dataset, write conversion scripts, create the TypeScript trainer, and add evaluation.
Review the generated code and changes.
The agent runs pnpm dev, and Arkor Studio opens locally at localhost:4000.
Click Run Training, monitor the loss and checkpoints, test the trained adapter, and deploy the result.
For example, you can give your coding agent a prompt like:
Use Arkor to fine-tune a model that rewrites rough drafts as tweets in my style: https://github.com/arkorlab/arkor Find or prepare a suitable dataset, create the TypeScript training workflow, and tell me when it is ready to review in Arkor Studio.
Arkor is not intended to be a magical prompt-to-model black box. It is a developer-controlled workflow where coding agents can handle much of the setup, while the training code, data transformations, evaluation, and final decisions remain visible and reviewable.
We’re especially interested in feedback on:
Whether the coding-agent workflow feels intuitive
Which parts of fine-tuning still feel unclear or intimidating
What you would need before using Arkor for a production model
Which models, datasets, and deployment workflows we should support next
Thanks for checking out Arkor. We’ll be here throughout the launch and would love to hear what you try building. 🙏
Speaking as someone without an ML background — fine-tuning a model for a specific task was never even on my radar. It sounded like something that needed a research team, a GPU cluster, and a lot of Python, so I'd just filed it under "not for me." Reading how Arkor works, it suddenly feels within reach.
The reason it clicks: I run a lot of automated workflows every day, all on general-purpose models. They work, but they're not tuned for my specific tasks. Being able to fine-tune a small dedicated model per flow — without touching any infra — could genuinely push the output quality further.
Curious: for someone with a handful of narrow, repetitive workflows, when does a fine-tuned model actually start to beat just prompting a bigger general model well?
Hi, it is really nice to have a chance to try for free as alpha users, and it is really useful product for people care about percision and latency.
I have a question, once we trained our model, how do connect to hugging face to publish our model?
Sincerely, Inky
The 'tell Claude Code what the model is for and it preps the dataset' handoff is the part people usually stall on, so nailing that is smart. When I hit Run Training in the local studio, is the GPU compute actually local to my machine, or does it provision remote GPUs behind that UI — and does the deployed OpenAI-compatible endpoint live self-hosted or on Arkor's side? Also curious how Claude Code/Codex actually drives the dataset prep: an MCP server, a CLI it shells out to, or just file conventions in the project?
Fine-tuning open-weight models in TypeScript is kind of unusual approach, most tooling for this is Python heavy. Makes it much more accesible for JS devs though. What models are supported so far?
the eval/leakage question below is the big one, but there's a second thing nagging me: this workflow lives inside your existing repo, with the coding agent preparing the dataset alongside the code. does that mean the actual training examples end up committed to git history by default? if someone points this at a real support-ticket or user-message dataset to fine-tune on, that's real customer data now living in version control unless you go out of your way to gitignore it - feels like an easy footgun for anyone who isn't already thinking about data handling.
Pushing a fine-tune through without touching a single Python script felt weirdly good. The handoff between Claude Code preparing the data and Arkor just clicking "Run Training" was exactly the kind of low-friction flow I keep wishing existed.
curious about the "no ML expertise required" claim specifically - when Claude Code or Codex picks the hyperparameters and dataset splits, how do you know the fine-tune actually improved anything vs just memorized the training set? that eval step usually needs someone who knows what they're looking at, not just infra automation.
Arkor makes the fine-tuning workflow feel much closer to normal app development, which is exactly where this has been too painful for most dev teams.
The part I’d want most visibility into is the dataset and eval step. If the coding agent prepares the data and trainer, does Arkor Studio surface checks for train/val leakage, near-duplicate examples, or weak eval coverage before the GPU time gets spent?
Looks super cool guys! Curious to know though, how much control do users have when choosing which llm they can fine tune?
The dataset-prep step is where I'd be nervous, since that's exactly the kind of thing a coding agent will happily do and happily get subtly wrong a bad train/val split, near duplicate examples leaking across the two, or silently deduping less than it should. None of that shows up as an error, it just shows up later as a model that looks great in Arkor Studio's loss curve and then underperforms in the real world.
Does Arkor Studio do anything to catch that automatically flag overlap between train/val sets, or run a sanity check before burning GPU time on the actual training run? That's the piece I'd want visible before trusting the agent-driven prep step for something going to production.
About Arkor on Product Hunt
“Fine-tune and Deploy Open-weight Models in TypeScript”
Arkor launched on Product Hunt on July 22nd, 2026 and earned 149 upvotes and 25 comments, placing #11 on the daily leaderboard. Start a real LLM training in 10 minutes. Tell Claude Code or Codex what the model is for; they prepare the datasets and create the training project. Click "Run Training" in the local studio. Arkor runs the training, and deploys the trained model as an OpenAI-compatible API. Think Next.js and Vercel for fine-tuning: code you can review, infrastructure you do not have to manage, and a model your app can call. No GPU setup. No ML expertise required. No Python training code to write.
Arkor was featured in Open Source (68.6k followers), Developer Tools (516.2k followers), GitHub (41.3k followers) and Development (6k followers) on Product Hunt. Together, these topics include over 119.4k products, making this a competitive space to launch in.
Who hunted Arkor?
Arkor was hunted by Hina. A “hunter” on Product Hunt is the community member who submits a product to the platform — uploading the images, the link, and tagging the makers behind it. Hunters typically write the first comment explaining why a product is worth attention, and their followers are notified the moment they post. Around 79% of featured launches on Product Hunt are self-hunted by their makers, but a well-known hunter still acts as a signal of quality to the rest of the community. See the full all-time top hunters leaderboard to discover who is shaping the Product Hunt ecosystem.
Want to see how Arkor stacked up against nearby launches in real time? Check out the live launch dashboard for upvote speed charts, proximity comparisons, and more analytics.
Hina
Hi Product Hunt! 👋
I’m Hina, one of the makers of Arkor.
We built Arkor because we kept running into the same gap: adding an AI feature to an application is easy, but adapting a model to a specific task still feels like joining an ML infrastructure team.
The moment you want to fine-tune, you often end up maintaining a separate Python project, converting datasets into unfamiliar formats, provisioning GPUs, and moving between scripts, notebooks, and dashboards.
We wanted model development to feel more like ordinary software development.
With Arkor, the workflow starts inside your existing repository:
Tell Claude Code or Codex what behavior you want from the model.
Your coding agent can find or prepare a dataset, write conversion scripts, create the TypeScript trainer, and add evaluation.
Review the generated code and changes.
The agent runs pnpm dev, and Arkor Studio opens locally at localhost:4000.
Click Run Training, monitor the loss and checkpoints, test the trained adapter, and deploy the result.
For example, you can give your coding agent a prompt like:
Arkor is not intended to be a magical prompt-to-model black box. It is a developer-controlled workflow where coding agents can handle much of the setup, while the training code, data transformations, evaluation, and final decisions remain visible and reviewable.
We’re especially interested in feedback on:
Whether the coding-agent workflow feels intuitive
Which parts of fine-tuning still feel unclear or intimidating
What you would need before using Arkor for a production model
Which models, datasets, and deployment workflows we should support next
Thanks for checking out Arkor. We’ll be here throughout the launch and would love to hear what you try building. 🙏