本网页来自:https://mp.weixin.qq.com/s/ir5xr2c84QRcKTRwpovecg,如果您不愿意显示在本站,请联系13951431913#139.com删除。
如果你正在找一个低门槛的多模态模型平台,个人学习、Demo 验证、轻量创作场景,Agnes AI这个平台可以先收藏一下。
Agnes AI 是什么?
Agnes AI 背后是 Sapiens AI,定位是做多模态 AI 模型和基础设施。
官方的愿景可以概括成一句话:让世界级 AI 人人可用。
这句话听起来有点大,但落到开发者这边,其实就是几个很现实的问题:
- 我能不能很快接入?
- 我能不能用一个 API 调文本、图片、视频?
- 我能不能先低成本试一试,而不是一上来就被复杂平台劝退?
Agnes AI 目前给出的答案是:可以。
它能做什么?
Agnes AI API 提供的是一套统一的多模态模型服务,常见能力包括:
- AI 聊天和文本生成
- 逻辑推理和内容理解
- 文生图、图像编辑和风格化处理
- 图生视频、文生视频
- 音视频同步生成
- Agent 工具和自动化工作流
- 创意内容生成和多模态交互应用
官网入口:https://agnes-ai.com/
目前平台里的模型覆盖文本、图像、视频等方向:
准备工作:先拿到 API Key
API 平台地址:https://platform.agnes-ai.com/
下面我用 Postman 演示三类模型:文本、图像、视频的使用。
体验 1:文本模型 Agnes 2.0 Flash
先试最基础的文本生成模型:agnes-2.0-flash。
在 Postman 里配置好刚刚创建的密钥:
请求参数如下:
{”model”: ”agnes-2.0-flash”,”messages”: [{”role”: ”user”,”content”: ”以“早”字为题,写一篇不少于500字的文章,文体不限.”}],”temperature”: 0.7,”max_tokens”: 2048}
生成结果:
体验 2:图像模型 agnes-image-2.1-flash
请求参数:
{”model”: ”agnes-image-2.1-flash”,”prompt”: ”未来城市日落,电影感。”,”size”: ”720x1280”,”extra_body”: {”image”: [],”response_format”: ”url”}}
生成结果:
这个能力比较适合用来做:
- 产品图生成
- 创意海报草图
- 人物图像生成
- 社交媒体配图
- 图片风格化和编辑
如果你平时需要给文章、短视频、产品 Demo 快速配图,这类模型的价值就很明显:先把方向跑出来,再决定要不要继续精修。
体验 3:视频模型 Agnes Video V2.0
它的调用方式和文本、图片略有不同,需要分两步走:
第一步:创建视频生成任务
请求参数:
{”model”: ”agnes-video-v2.0”,”prompt”: ”帮我生成一个视频:人物面带职业微笑,右手撩一下刘海往镜头正前方走两步,右手叉腰侧身展示套装,再后退两步右手挥手打招呼,拿手机的左手保持原样不变,整套动作连贯,自然流畅,高清融合。”,”frame_rate”: 60}
请求成功后,会返回一个 video_id:
{”id”: ”task_c7a5pKSIdEkNkbldLxT7HcW6NHatdRrN”,”video_id”: ”video_bGl0ZWxsbTpjdXN0b21fbGxtX3Byb3ZpZGVyOm9wZW5haTttb2RlbF9pZDphZ25lcy12aWRlby12Mi4wO3ZpZGVvX2lkOnZpZGVvXzdiNTE2NjU4Nzg2NDRkMTRiMjFkODEwMWE5ZDI2MDg1”,”task_id”: ”task_c7a5pKSIdEkNkbldLxT7HcW6NHatdRrN”,”object”: ”video”,”model”: ”agnes-video-v2.0”,”status”: ”queued”,”progress”: 0,”created_at”: 1783578316,”seconds”: ”2.0”,”size”: ”1088x832”}
第二步:查询视频结果
拿到第一步返回的 video_id 后,再调用查询接口:
请求完成后,会返回生成视频的 URL:
视频模型的使用场景很清晰,比如产品展示、人物动作演示、短视频素材、营销创意草稿等。它不一定一上来就能替代专业视频制作,但用来快速验证创意,效率会比从零制作高很多。
感兴趣的话,可以从官方文档获取更多信息https://agnes-ai.com/zh-Hans/docs/overview
如果你最近正好想找一个免费或低成本的大模型平台做实验,推荐你试试agnes-ai。
This webpage is from: https://mp.weixin.qq.com/s/ir5xr2c84QRcKTRwpovecg. If you do not want it displayed on this site, please contact 13951431913#139.com for removal.
If you are looking for a low-barrier multimodal AI model platform for personal learning, demo validation, and lightweight creative projects, I'd recommend bookmarking Agnes AI first.
What is Agnes AI?
Agnes AI is backed by Sapiens AI, positioned as a provider of multimodal AI models and infrastructure.
Their official vision can be summarized in one sentence: making world-class AI accessible to everyone.
This may sound grand, but from a developer's perspective, it boils down to a few practical questions:
Can I integrate quickly?
Can I use a single API to call text, image, and video models?
Can I try it at low cost first, without being discouraged by complex platforms upfront?
Agnes AI's current answer is: yes.
What can it do?
The Agnes AI API provides a unified suite of multimodal model services, with common capabilities including:
AI chat and text generation
Logical reasoning and content understanding
Text-to-image, image editing, and style transfer
Image-to-video and text-to-video
Audio-video synchronized generation
Agent tools and automated workflows
Creative content generation and multimodal interactive applications
Official website: https://agnes-ai.com/
The models currently available on the platform cover text, image, video, and other areas:
Preparation: First, obtain an API Key.
API platform address: https://platform.agnes-ai.com/
Below, I'll demonstrate three types of models — text, image, and video — using Postman.
Experience 1: Text Model Agnes 2.0 Flash
Let's start with the most basic text generation model: agnes-2.0-flash.
Configure the newly created key in Postman:
Request parameters:
Generation result:
Experience 2: Image Model agnes-image-2.1-flash
Request parameters:
Generation result:
This capability is well-suited for:
Product image generation
Creative poster sketches
Character image generation
Social media graphics
Image stylization and editing
If you frequently need to quickly pair images for articles, short videos, or product demos, the value of this type of model becomes clear: run with the direction first, then decide whether to continue refining.
Experience 3: Video Model Agnes Video V2.0
Its invocation method differs slightly from text and image — it requires a two-step process:
Step 1: Create a video generation task
Request parameters:
Upon successful request, a `video_id` is returned:
Step 2: Query the video result
After obtaining the `video_id` from Step 1, call the query API:
After the request completes, the generated video URL is returned:
The use cases for the video model are straightforward — product showcases, character action demonstrations, short video素材, marketing creative drafts, etc. It may not replace professional video production right away, but for quickly validating ideas, it's far more efficient than starting from scratch.
If interested, you can find more information in the official documentation: https://agnes-ai.com/zh-Hans/docs/overview
If you've been looking for a free or low-cost large model platform for experiments lately, I recommend trying agnes-ai.