Right now, a lot of people are probably eager to experience just how powerful AI agent tools like OpenClaw and Hermes Agent are, but the vast majority of APIs require payment, and even when free, they’re usually very slow. Recently, I came across a service called Anges AI from Singapore, where many of its current models are available for free via API calls, with no usage limits for now and no credit card required.
More importantly, it’s not just text—there are also models for generating images and videos. I tested it during off-peak hours, and the speed was pretty good. Those who need it can give it a try.
Completely free AI API! Supports AI agents, image and video generation—no credit card required.
Anges AI offers the following models that can be integrated for free via API:
- Agnes-2.0-Flash (Text, AI Agents, Code Writing)
- Agnes-Image-2.0-Flash, Agnes-Image-2.1-Flash (image generation)
- Agnes-Video-2.0 (Video Generation)
The official team also shared benchmark scores: on the ClawEval general leaderboard, its text model ranks ahead of GPT 5.4. On the AA image editing leaderboard, its image model scores close to Nano Banana. On the AA image-to-video leaderboard, its video model only trails Seedance 1.5 Pro by a small margin. I tested the image and video generation quality, and it is indeed pretty good:
Applying for API access is also very simple—just register an account. No credit card is required.
Go to the Anges AI official website.
Click the link above to visit the Anges AI official website, then register an account first. Google and GitHub login are also supported:
After logging in, open the API Keys page and click “Create new secret key” at the top to create an API key:
Enter secret key name:
You will then obtain the API Key:
Going forward, you can also copy this API Key in the backend at any time:
Next, it depends on where you want to use it. Below, I’ll demonstrate with Hermes Agent. For the API Base URL, enter https://apihub.agnes-ai.com. You can find the remaining settings in the documentation at the top of the admin panel. However, the documentation page often shows errors, so you’ll need to keep refreshing until it displays properly.
You can configure the API directly in the Hermes Agent desktop app, or through the terminal (the URL below was entered incorrectly when I took the screenshot; please ignore it):
After configuring the API, these models will appear in the model menu. Note that if you want to use models that generate images or videos, you don’t need to switch to Image or Video; just ask the text model to call them:
Tested the Anges 2.0 Flash text model: it successfully completed the task I requested—gathering the day’s new articles from the PCADV website and saving them as a CSV file on the desktop:
Image generation was also successful, the quality is decent, and the generation speed is fast—it was completed within 1 minute:
The video has also been generated successfully. Please wait about 2 minutes:
However, note that I’m not sure whether it’s because Agnes-2.0-Flash isn’t smart enough or for some other reason, but whenever I use it to request video generation or query already-generated videos, I get an “invalid token” error. Switching to Gemini 3.1 Flash Lite worked fine for me, though. If you run into the same issue, I’d suggest trying a different model:
Source: KOCPC Chinese