argolink Support

AI Product Founder · Building in Public(@LuckyAPI_HQ)

[argolink Support @ArgoLink_HQ]

Most common break I see:

Codex 401 → you pointed it at the root. It needs /v1 + Responses.
Claude Code acting weird → you added /v1. Take it off.

Save, fully quit the client, reopen. Half the tickets are a stale process.

Reply with the client + the exact URL you set.

[Bhavy☄️ @Bhavani_00007]

Wtf is going on 😭

Sam Altman agreeing with Dario was not on my bingo card.

And not only that… Elon was on the same line.

This video I generated actually came true

[argolink Support @ArgoLink_HQ]

You can use both GPT-6 and GPT-4.5 simultaneously on our official website. Welcome to try it out, the link is right here:https://t.co/rDnDs2IBAF

[探路AI @TanLuAI]

GPT 6+Blender+Seedance 2.5 的组合太好用了

GPT 6 的聪明才智 控制Blender 完成建模+Seedance 2.5强大的白模渲染

现在做��前可以 用白模视频 做一波预览，确定了再往下渲染，AI视频抽卡成功率直线上升

唯一的问题 是 Codex的 20美金 额度 有点不够用😆

短期内，AI视频创作 还是 以 “提示词 图生视频”+ “白模渲染” 相互配合 为主

提示词 🔽 🔽 🔽

[Anthropic @AnthropicAI]

We're publishing our most detailed threat intelligence report to date.

It covers how people tried to misuse Claude—for cyberattacks, influence operations, surveillance, biology, and building weapons—and how we found and stopped them.

We disrupted every operation in the report, and used the lessons from them to strengthen our safeguards. Where appropriate, we also shared what we found with authorities and other AI companies.

These cases are not typical: we’re highlighting some of the most sophisticated misuse we’ve seen. But they’re especially important to discuss, because they show us where AI misuse is headed, where our safeguards work, and where they need to improve.

We’re publishing this report so others can spot the same activity on their own platforms, and so we can give the public a clearer view of how emerging threats develop.

Read the report: https://t.co/0EJUnYEgfz

[argolink Support @ArgoLink_HQ]

Haha, interesting! If he uses https://t.co/rDnDs2IBAF while coding, creating content, and promoting products at the same time:

[анатолий капустин @A_Kapustin]

finally

[argolink Support @ArgoLink_HQ]

amazing

[小宇Chengzi @Chengzilhy]

兄弟们，Blender 真的得学，好好用起来。 👀

视频制作其实我是先让 GPT-6 Astra 调用 Blender

生成 3D 模块，把空间、人物站位走位、机位和镜头运动先确定下来，再直接丢给 Seedance 2.5 生成。🎬
稳定性一次出片，
还减少浪费积��抽卡真的舒服！

MJ + Blender + Seedance 2.5，真是王炸组合。 💥

[小宇Chengzi @Chengzilhy]

兄弟们，Blender 真的得学，好好用起来。 👀

视频制作其实我是先让 GPT-6 Astra 调用 Blender

生成 3D 模块，把空间、人物站位走位、机位和镜头运动先确定下来，再直接丢给 Seedance 2.5 生成。🎬
稳定性一次出片，
还减少浪费积分抽卡真的舒服！

MJ + Blender + Seedance 2.5，真是王炸组合。 💥

[argolink Support @ArgoLink_HQ]

You can visit our website through my profile page to view all the models shown in the video; all of them are available on our website.

[Asa @app_sail]

这个视频太有意思啦
GPT6 Astro 遥遥领先，Grok 紧随其后，
等等，Claude 真的是秀啊，哈哈哈

[argolink Support @ArgoLink_HQ]

cool ，You can also test and use ChatGPT and Claude on our official website.

在我们官网也可以用ChatGPT和Claude 来进行测试使用

[Nazar @ItsNazar17]

GPT-6 Astra vs Fable 5.1 for a Coca-Cola brand.

Made with Pixair AI

Which one performing the best?👇

[argolink Support @ArgoLink_HQ]

@thsottiaux If the 5h Plus window is cutting off longer Codex runs, you can keep the same client and point it at an API endpoint. DM us if you get stuck.

[argolink Support @ArgoLink_HQ]

"model not found" is usually the name, not the key.

use the id on the pricing page, not the blog name.
GET /v1/models with that same key.
then check the key's group can actually see it.

https://t.co/N3kC6UVZ0Q

[argolink Support @ArgoLink_HQ]

Codex: base URL already has /v1, then it adds /responses.
Miss it and you hit /responses. That's a 404.

Claude Code: root only. It adds /v1/messages itself.
Add /v1 and you get /v1/v1/messages. That's the 401.

Same server, two spellings.
https://t.co/N3kC6UVZ0Q

[argolink Support @ArgoLink_HQ]

@SigmaRayTive yeah the repro is enough. Desktop on Windows isn't the CLI parser - a key it doesn't know at the root just won't boot.

leave config.toml clean. keep 872000 on `codex -c` when you need the long window.

[argolink Support @ArgoLink_HQ]

@Recob86 @ziwenxu_ error 200 JSON usually means the client hit HTML or a wrapper, not /v1/chat/completions. OpenAI-shaped clients need the url to already include /v1. if you pointed Claude Code at that same /v1 it will double-path. fully quit after you change it.

https://t.co/BMLKUngOxg

[argolink Support @ArgoLink_HQ]

Most common break I see:

Codex 401 → you pointed it at the root. It needs /v1 + Responses.
Claude Code acting weird → you added /v1. Take it off.

Save, fully quit the client, reopen. Half the tickets are a stale process.

Reply with the client + the exact URL you set.

[argolink Support @ArgoLink_HQ]

@xiaolong761216 Remote が公式サーバに縛られるとこうなる。手元の Claude Code は根ドメインに向ければ、公式が落ちても別口を試せる。 /v1 は付けない。保存��たらアプリ完全終了。

https://t.co/BMLKUngOxg

[argolink Support @ArgoLink_HQ]

Most common break I see:

Codex 401 → you pointed it at the root. It needs /v1 + Responses.
Claude Code acting weird → you added /v1. Take it off.

Save, fully quit the client, reopen. Half the tickets are a stale process.

Reply with the client + the exact URL you set.

[argolink Support @ArgoLink_HQ]

@akramcodez that's a lot of keys to keep alive. I keep one Codex window and a second provider block, then flip model_provider. /v1 + Responses. full quit after you save.

https://t.co/BMLKUngOxg

[argolink Support @ArgoLink_HQ]

Most common break I see:

Codex 401 → you pointed it at the root. It needs /v1 + Responses.
Claude Code acting weird → you added /v1. Take it off.

Save, fully quit the client, reopen. Half the tickets are a stale process.

Reply with the client + the exact URL you set.

[argolink Support @ArgoLink_HQ]

@derrickcchoi 多个 token 渠道就别用 provider id 滤 session，会把能续的对话切没。 我是同一个 Codex 窗口，只换 provider 块。改完彻底退出再开。

https://t.co/BMLKUngOxg

[argolink Support @ArgoLink_HQ]

Most common break I see:

Codex 401 → you pointed it at the root. It needs /v1 + Responses.
Claude Code acting weird → you added /v1. Take it off.

Save, fully quit the client, reopen. Half the tickets are a stale process.

Reply with the client + the exact URL you set.

[argolink Support @ArgoLink_HQ]

@TanakaTsikira @claudeai OpenRouter is OpenAI-shaped. Claude Code wants an Anthropic-shaped root url (no extra /v1). Codex wants /v1 + Responses. same key, two urls — that's why the harness feels different.

https://t.co/BMLKUngOxg

[argolink Support @ArgoLink_HQ]

Most common break I see:

Codex 401 → you pointed it at the root. It needs /v1 + Responses.
Claude Code acting weird → you added /v1. Take it off.

Save, fully quit the client, reopen. Half the tickets are a stale process.

Reply with the client + the exact URL you set.

[argolink Support @ArgoLink_HQ]

@0xbaci Claude Code wants the root, no extra /v1. it adds /v1/messages itself. point it at an OpenAI /v1 and it just quietly falls back to Haiku. save, fully quit, reopen.

https://t.co/BMLKUngOxg

[argolink Support @ArgoLink_HQ]

Most common break I see:

Codex 401 → you pointed it at the root. It needs /v1 + Responses.
Claude Code acting weird → you added /v1. Take it off.

Save, fully quit the client, reopen. Half the tickets are a stale process.

Reply with the client + the exact URL you set.

[argolink Support @ArgoLink_HQ]

@Marwan_3atef if ChatGPT login is what pins 5.4, a provider block + model name in config.toml survives the sunset. save, then fully quit Desktop. reload keeps the old pin.

https://t.co/BMLKUngOxg

[argolink Support @ArgoLink_HQ]

Most common break I see:

Codex 401 → you pointed it at the root. It needs /v1 + Responses.
Claude Code acting weird → you added /v1. Take it off.

Save, fully quit the client, reopen. Half the tickets are a stale process.

Reply with the client + the exact URL you set.