Short answer: Agnes AI is a Singapore-based multimodal API platform by Sapiens AI that offers free access to text, image, and video models with a 512K context window, OpenAI-compatible endpoints, and no credit card requirement.
- Agnes AI provides free access to text, image, and video models with no credit card or waitlist required.
- The flagship text model agnes-2.0-flash supports a 512K context window, tool calling, streaming, and image understanding.
- API endpoints are fully OpenAI-compatible: swap the base URL and existing code works without changes.
- Pricing is documented as $0 per 1M input and output tokens, though the official docs carry a caveat that terms may change.
- The platform also ships image generation (Agnes Image 2.0/2.1 Flash), video (Agnes Video V2.0), and a no-code creative studio called Pavo.
Table of contents
- What is Agnes AI?
- Models and capabilities
- Pricing and access
- OpenAI compatibility
- How Agnes AI compares to other free tiers
- Risks and caveats
- Beyond the API: the Agnes ecosystem
- Should you use Agnes AI?
- Conclusion
- Frequently asked questions
- References
What is Agnes AI?
Agnes AI is the consumer-facing brand of Sapiens AI, a Singapore-based company that positions itself as "Singapore's homegrown model company." The platform launched its permanent free tier on June 1, 2026, offering unrestricted access to text, image, and video generation through a single API key. A secondary report from OpenAI Hub published on the launch date confirmed the free tier was announced as universal and permanent from day one, covering all three modalities without requiring a credit card or waiting for approval.
The company's public narrative emphasizes that Agnes AI is built for developers who want to prototype multimodal applications without paying upfront. The brand also extends beyond the API: Sapiens AI has released Agnes Code, a desktop application for coding workflows, and Pavo, a no-code creative studio for image and video generation. The Agnes AI homepage lists these products alongside the model catalog, signaling that the company is building a full developer and creator toolkit, not just an API endpoint.
Models and capabilities
The flagship text model is Agnes 2.0 Flash (model ID: agnes-2.0-flash). According to the official documentation, it supports a 512K context window, streaming responses, tool calling, image-url input, and image understanding. The docs also list agent workflows and coding tasks among its strengths. On the Claw-Eval general leaderboard, Agnes 2.0 Flash ranks at position 9 with a Pass@3 score of 60.9%.
For image generation, the platform offers Agnes Image 2.0 Flash and the newer Agnes Image 2.1 Flash, supporting both text-to-image and image-to-image editing. The video model, Agnes Video V2.0, rounds out the multimodal stack. These model families are publicly listed in the official GitHub repository, which also documents the OpenAI-style workflow and includes a Python quickstart example.
The model catalog confirms that Sapiens AI is pursuing an omni-modal strategy: one API, three generation types. That is unusual for a free tier. Most providers that offer free access specialize in one modality -- typically text or chat -- and charge for image or video generation. Agnes AI bundles all three under the same free access policy.
Pricing and access
The current pricing, as documented on the Agnes 2.0 Flash docs page, lists both input and output tokens at $0 per 1 million tokens. Authentication is via a Bearer API key obtained through the Agnes AI developer portal. There is no credit card requirement and no waitlist. However, the documentation carries an explicit statutory caveat: "Model availability, rate limits, pricing, and quota rules may change over time." This means the free tier is free as of the documentation snapshot, not guaranteed indefinitely.
Coverage from e27 on July 13, 2026 reported that Agnes 2.5 Flash was gray-released as an uncapped free tier aimed specifically at coding and agentic tasks, suggesting the company is actively iterating on its free access policy even as it expands toward a Pro flagship model.
TechTimes reported that the platform hit 4 trillion free API calls within the first week of the permanent free access announcement. The same outlet covered the simultaneous launch of Pavo, the no-code creative studio, positioning it as part of a broader developer and creator ecosystem.
OpenAI compatibility
The main selling point for developers is OpenAI compatibility. According to the official docs overview, the base URL is https://apihub.agnes-ai.com/v1, and the endpoint structure mirrors OpenAI's /v1/chat/completions format. This means existing code that uses the OpenAI Python or Node.js SDK can switch providers by changing a single configuration value -- the base URL.
Community developers have already built integration wrappers around this pattern. A community skill on GitHub shows a practical workflow using persistent authentication and the OpenAI-style client, confirming that the ecosystem is already building tooling on top of Agnes AI's API shape.
How Agnes AI compares to other free tiers
| Provider | Free Tier | Context | Multimodal | OpenAI Compatible | Notable Caveat |
|---|---|---|---|---|---|
| Agnes AI | Free text/image/video | 512K+ | Yes | Yes | Docs say pricing/quota may change |
| Groq | Free/usage-based | Model-dependent | Mainly text/vision | Partial | Rate limits vary by model |
| Google AI Studio | Free tier | Often 1M+ | Text, image, video | Growing | Rate limits; policy shifts |
| OpenRouter | Free routing | Model-dependent | Limited | Yes | Free models deprioritized |
| DeepSeek | Free/quota | 64K-128K | Text-focused | Partial | UI mainly Chinese-facing |
This comparison is based on publicly documented free-tier terms as of July 2026. Actual availability and limits should be confirmed at the time of integration, as the table above reflects each provider's stated terms, not independent load testing.
Risks and caveats
The most important risk with any free API is stability. Fintech Singapore's coverage of the 2.5 Flash release noted that the uncapped coding tier was still in gray-release at the time of reporting, which suggests the platform is actively managing capacity and may throttle or modify access before settling on a final policy.
There is no verified public SLA for Agnes AI's free tier. Production teams should assume that free access carries soft rate limits and no uptime guarantee. The official docs warn that quota rules may change, and the rapid product iterations -- 2.0 Flash in June, 2.5 Flash in July, a Pro model teased -- indicate a company still finding its footing on the infrastructure side.
Benchmark claims on the Claw-Eval and Artificial Analysis leaderboards are self-reported or lab-published. They should be read as directional signals rather than independent validation. For developers evaluating Agnes AI for production workloads, the practical test is always your own prompts against your own use case.
Beyond the API: the Agnes ecosystem
Sapiens AI is not just building an API. The company has released Agnes Code, a desktop application for coding workflows, and Pavo, a no-code creative studio for image and video generation. According to e27's July 13 report, the Agnes Code app was unveiled alongside the 2.5 Flash model, and a flagship Agnes 2.5 Pro model was teased at the same event. The Pavo studio was introduced earlier, in late June, alongside the 4 trillion API calls milestone.
This ecosystem play -- API plus desktop app plus no-code studio -- mirrors the strategy that major AI companies used to lock in developers: give away the tools, then monetize at scale. Whether Agnes AI follows the same trajectory remains to be seen, but the pattern is familiar.
Should you use Agnes AI?
Agnes AI is worth testing if you need a free, multimodal API that works with existing OpenAI code. The 512K context window on the text model is strong for a free tier, and the ability to generate images and video through the same endpoint is unusual. If you are prototyping a multimodal agent, building a creative tool, or evaluating models before committing to a paid provider, Agnes AI offers a low-friction entry point.
The main caution is that free access terms can change. The documentation explicitly warns that pricing and quota rules may shift. For production workloads that require guaranteed uptime and SLA coverage, you will want to verify the current state of the free tier at publishing time, rather than relying on documentation from weeks or months ago. The rapid growth trajectory -- 4 trillion calls in one week -- suggests demand is real, which could pressure capacity and accelerate policy changes.
If you are evaluating free API tiers for your next project, the better move is to run your actual workloads against Agnes AI's endpoints before making architectural decisions. The OpenAI-compatible shape means the cost of switching is minimal, so testing is cheap and the signal is reliable.
Conclusion
Agnes AI offers a genuinely free, multimodal API with 512K context and OpenAI compatibility, making it one of the lowest-friction entry points for developers building text, image, and video applications. The free tier is real today, but its longevity depends on how Sapiens AI manages scaling and monetization.
Frequently asked questions
Is Agnes AI really free forever?
The free tier is free as documented today, but the official Agnes AI docs carry an explicit caveat that "model availability, rate limits, pricing, and quota rules may change over time." There is no guarantee of indefinite free access.
How do I migrate my OpenAI code to Agnes AI?
Change the base URL to https://apihub.agnes-ai.com/v1 and set your API key as a Bearer token. The endpoint structure matches OpenAI's /v1/chat/completions, so no code changes are needed beyond configuration.
What models does Agnes AI offer?
Agnes AI offers agnes-2.0-flash (text, 512K context), agnes-image-2.0-flash and agnes-image-2.1-flash (image generation/editing), and agnes-video-v2.0 (video generation). Agnes 2.5 Flash was also gray-released for coding and agentic tasks.
Where does Agnes AI rank on benchmarks?
On Claw-Eval's general leaderboard, agnes-2.0-flash ranks at position 9 with a Pass@3 score of 60.9%. Agnes AI also reports leadership on Artificial Analysis image and image-to-video leaderboards, though these are self-reported claims.
Who runs Agnes AI?
Agnes AI is the consumer brand of Sapiens AI, a Singapore-based company. The company markets itself as "Singapore's homegrown model company" and has released a suite of products including the API, Agnes Code desktop app, and Pavo creative studio.
References
- Agnes AI homepage -- product narrative and model catalog
- Agnes AI docs overview -- OpenAI compatibility and base URL
- Agnes 2.0 Flash documentation -- specs, pricing, and benchmark claims
- AgnesAI-Labs/AgnesAI-Models GitHub repository -- model catalog and quickstart
- TechTimes -- 4 trillion free API calls and Pavo studio launch
- e27 -- Agnes 2.5 Flash, Agnes Code desktop app, and Pro model tease
- Fintech Singapore -- Agnes 2.5 Flash agentic coding coverage