# ⚡ Free coding agent that writes its own notes (Prime Agent)

**URL:** <https://onehack.st/t/free-coding-agent-that-writes-its-own-notes-prime-agent/327384>\
**Category:** Tools & Scripts\
**Tags:** programming, tips-tricks, ai\
**Created:** [November 20, 2026, 6:55pm UTC](https://onehack.st/t/free-coding-agent-that-writes-its-own-notes-prime-agent/327384 "2026-11-20T18:55:13Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![SRZ](https://onehack.st/user_avatar/onehack.st/srz/32/146132_2.png) [@SRZ](https://onehack.st/u/SRZ)\
**Post date:** [November 20, 2026, 6:55pm UTC](https://onehack.st/t/free-coding-agent-that-writes-its-own-notes-prime-agent/327384/1 "2026-11-20T18:55:14Z")

</div>

![PRIME](https://onehack.st/uploads/default/original/3X/8/7/87d14530b16d4bf65be5f1b9ffd419101a9e3ac8.jpeg)

⚡ A free coding agent that writes its own notes — so it gets better the more you use it.

[Prime Agent](https://github.com/PrimeIntellect-ai/prime-agent) is open source (MIT, ~22.8k stars). You bring a free model key; it does the work.

🗝 **1 · Run one line**  
[Ori](https://openrouter.ai/ori/harness/prime-agent) → pick a free model → one command. No card.

💻 **2 · Give it a small real task**  
“build a tiny script” · “fix this error” · “explain this file”.  
Its only tool is a small Python workspace — it writes code to do the job.

🧩 **3 · It hires helpers**  
It opens copies of itself for parts of the work. Insiders: `rlm.spawn()` hands back a job ticket first, the answer later.

📝 **4 · It saves the lesson**  
After the run it writes notes and skills into its own memory — a reversible edit. That is the “self-improving” part.

```auto
   task
     │
     ▼
   code ──► helpers
     │
     ▼
   memory
     │
     ↺──► next run, better

```

🔁 **5 · The cheeky bit**

A four-file MIT cousin, [telos](https://github.com/adntaha/telos), goes further.  
It rewrites its own source and restarts mid-chat — it even writes its own first system prompt.  
Keep it in a throwaway box.

Start free: one command, one free key, one small real task.

> **What it is, every free door, the bolt-ons — and the honest bits**
>
> 🧠 **What it is — insider words with plain meanings**  
> • [Prime Agent](https://github.com/PrimeIntellect-ai/prime-agent) — ~22.8k stars · [MIT](https://github.com/PrimeIntellect-ai/prime-agent/blob/main/LICENSE) — use it, change it, share it · [current release](https://github.com/PrimeIntellect-ai/prime-agent/releases)  
> • [the maker’s plain explanation](https://www.primeintellect.ai/blog/prime-agent) · [the paper](https://arxiv.org/html/2608.23552) — how the memory loop works  
> • insiders call the setup RLM. It copies itself into helpers (rlm.spawn) and saves what worked into notes and skills that outlive the chat — the continual harness. Each edit is logged and reversible.
> 
> 🗝 **Free doors — pick one**  
> • [Ori](https://openrouter.ai/ori/harness/prime-agent) — logs you in, no key to paste · [one-line install](https://openrouter.ai/labs/ori/install.sh)  
> • [OpenRouter free models](https://openrouter.ai/collections/free-models) — 50 calls a day at $0 · [limits](https://openrouter.ai/docs/faq) · [live provider table](https://llmperks.com/providers) · [llm-keypool](https://github.com/piyush-tyagi-13/llm-keypool) — free keys on rotation  
> • [Google AI Studio](https://aistudio.google.com/apikey) — big daily allowance · [Groq free](https://console.groq.com/settings/billing/plans) — very fast, short chats · [Ollama](https://docs.ollama.com/quickstart) — no account, no key, runs on your own machine · [Codespaces](https://github.com/features/codespaces) — 120 free hours in a browser  
> • [direct install](https://app.primeintellect.ai/prime-agent/install.sh) + [provider guide](https://getprimeagent.ai/guides/providers) — the hands-on route
> 
> 🧰 **Bolt-ons**  
> • [browser-use 1.0](https://github.com/browser-use/browser-use/releases/tag/1.0.0) — the same one-helper idea for web pages · [SkillLib](https://github.com/DistriAI/SkillLib) — 20,000+ skills · [qai.io skills](https://qai.io/skills) — 14,000+ more · [agentic-awesome-skills](https://github.com/sickn33/agentic-awesome-skills) — 3,000+ more  
> • [rat](https://github.com/MaximeRivest/rat) and [pythond](https://pythond.sh) — share one live Python space across programs · [marimo pair](https://marimo.io/blog/marimo-pair) — helpers linked by visible variable lines  
> • [Langfuse](https://github.com/langfuse/langfuse) — see every helper call · [E2B](https://sandboxindex.dev/) — a throwaway machine for risky code · [more sandbox options](https://upstash.com/blog/best-e2b-alternatives)  
> • [mengram](https://github.com/alibaizhanov/mengram) — memory that learns from mistakes · [primecord-agent](https://github.com/OpenCnid/primecord-agent) — always-on chat version · [A2A](https://github.com/a2aproject/A2A) + [orxhestra](https://github.com/nicolailassen/orxhestra) — helpers from other tools · [ACE](https://ace-agent.github.io/) — the playbook method behind the memory  
> • [community setup guide](https://cskwork.github.io/prime-agent-guide/) · [memory atlas](https://neoneye.github.io/agent-memory-atlas/systems/prime-agent/) — first-run fixes and field notes
> 
> 🔍 **Honest bits, with receipts**  
> • the 95.5% score you may have seen is a self-reported efficiency figure on an easier public set, not an official result — [plain explainer](https://thegenaimagazine.com/did-prime-agent-solve-arc-agi-3-95-percent-explained) · [its own run-by-run table](https://github.com/PrimeIntellect-ai/arc-agi-3-prime-agent) · [why the model matters more](https://www.engineeringmaxxing.com/veanors/papers/prime-agent.html)  
> • `/refine` edits notes and skills only · free keys suit short chats; long runs cost real money · the agent runs code with your permissions — use a throwaway folder.

Send one message. Watch it remember. 💬
