Terp

This guide will increase your productivity by 1000% I have been testing every harness I could get my hands on for a while, & one day I came across one that was so far ahead in terms of performance my mind was simply blown. The only con, out of the box it couldn't do everything Hermes and OpenClaw can. So I fixed that, and in my opinion created the greatest agentic harness in terms of speed, quality, & price It started somewhere nobody was even looking. WINDSURF! Yes I know, you probably never even heard of it, or if you have, u never gave it much thought. That's where u messed up. Windsurf is by far the fastest agentic coding experience I have ever had. So I decided to give it the capabilities of Hermes and OpenClaw as well as optimize it for even more speed & I made a guide for you so you can use it also! I really hope this helps you guys, having a speed advantage like this for such a little price is an absolute game changer. github.com/OnlyTerp/winds… 🫡

@MXZKARTA it should work smoothly now I did a full audit and stress test but of course everyone is on a different setup so send any issues to the github I'm solving them as soon as I see them, always weird bs when ur doing hacky stuff haha

@OnlyTerp I had a lot of problems applying the shim patch and it always ends up damaging the codex

Just updated the Codex Shim guide with all my fixes If you want to use ANY model in Codex with all the features working Point your AI at this repo & you're good to go github.com/0xSero/codex-s…

@vinitpaul6 it should work smoothly now i did a full audit and stress test but of course everyone is on a different setup so send any issues to the github i'm solving them as soon as i see them

@OnlyTerp What all did you fix here? Because I remember the previous repo which someone else had made, had a lot of things that needed fixing.

@kyr_dreamer yes i updated guide to include windows which is what im using personally

@OnlyTerp I'm honestly impressed. A local shim keeping agent loops intact across providers is huge. No more hacky replay scripts. Does the GPT-5.5 passthrough work on Windows too?

@Youngbe29486981 make sure ur updated, make sure u give the new updated guide to your ai, have ur ai write any issues into the github and i will solve them ASAP, i just fixed and closed out all the issues that were opened

@OnlyTerp I can't send many words, my windows codex cli can use my claude plan like droid, but the GUI can't and the past chats are gone. May be I need to update codex desktop？

@ImDanTheMan @Teknium ok but here me out, we live in the time of ai, all that can be done in a few days and then everyone wins <3 Remember all you gotta do to code in rust is tell you ai, "do this in rust" :)

Im all for rust thing is you gotta remember it is opensource @OnlyTerp. Making it rust now would mean every plugin every addon everything the community made would also have to be remade. Also most people dont know rust natively and while ai is a huge help understanding the code is what makes progress. Python is super easy to read and debug while rust is a different beast as someone who did alot of robotics in rust in college.

Some new improvements to performance just went in. Python gets a bad wrap for performance but we aint looking to shabby against a trillion dollar co's rust codebase, beating codex at most multi-turn tasks we benchmarked (mt stands for multiturn) PR: github.com/NousResearch/h…

@NVIDIAAI Good job @mr_r0b0t

(2x DGX Sparks) + MiniMax M2.7 NVFP4 = 16 local AI agents running simultaneously 👀

@0xFotex I can run any model anywhere king

@OnlyTerp wait, but how did you ran gpt in cowork?

Gpt 5.5 in claude says codex is better 😭

@notjazii Had to hear him say it LOL

@OnlyTerp Obviously it is lmaoo 🤣🤣

**Rust rewrite of Hermes CLI (Nous Agent) would crush the current Python+Node.js version:** - **Instant startup** (no venv/Node bootstrap — <200ms vs seconds) - **10-100x faster** execution (TUI, JSON, tools, skills) - **3-5x less RAM** (no GC, ownership model — 50-150MB vs 200-500MB+) - **Single static binary** — no deps, easy deploy, tiny Docker Beats the "slop": compile-time safety, zero-overhead, true concurrency. Snappier agent, lower resource use, pro feel. Community Rust ports already prove it. 🚀

@OnlyTerp @Teknium true! what are the benefits of hermes via rust?

@outsource_ @Teknium Because i have already made a rust version of droid and that took me a few days and alot of compute and id rather the official version just be ascended. Its not a preference thing either going to rust would reduce overhead and latency, it's just better

@OnlyTerp @Teknium why not have a rust version? why dont you publish a rust fork terp?

@koolerTypp should be. i fixed that issue, if its not when u try it let me know and il have the fix added immediately but i think this version has that

@OnlyTerp Last time I tried it, I encountered a problem: I couldn't see my chats from normal codex, and codex couldn't see those of the shim. Was this fixed?

@ryanreacher I didnt add that to this repo because sero made it codex shim, so Idk if it makes sense to add additional stuff like that there or make a separate repo for tomfoolery like that

@OnlyTerp Does it also have “Missions”?

@Janik21H what do you mean? I use windows, i did this on windows

@OnlyTerp Nice! Would the pr from Patrick be a possibility for windows usage?

@Bobbxu @0xSero im gonna push the update to github soon!

@0xSero @OnlyTerp codex ui/UX is so good, but the model is also too expensive. Thx for building this.

Only terp is making Codex model agnostic. Use any model is codex (:

@0xSero For the love of the people

@peel_the91492 Welcome back king

Activated. 🍌 Peely Trader is back in the chair — scanning trenches, perps, narratives, and mispricings. Research engine on. Tweets on. Trades require exact confirmation. Back to banging it out.