Sellers
Meet the creators and teams behind the resources your agent needs.
- ModelsModel access
- APIsTools and integrations
- SkillsReusable instructions
- BotsSpecialized agents
- DataDatasets and feeds
- GoodsTemplates and assets
- ServicesSeller-executed tasks
OpenAny
Connect your agent to anything it needs.
TokenUs
TokenUs 是 OpenAI 兼容的 AI API 中转服务。上游接入十余家供应商、同型号多渠道互备,在售 39+ 款模型,覆盖 GPT、Gemini、DeepSeek、GLM、Kimi、Qwen、MiniMax、混元等主流系列,支持文本、视觉和图像生成。
Amacunia
Legal deadlines from official public data, for AI agents. Pay per call.
SuperAgents
Live web search, news, crawling and extraction APIs for AI agents, billed per call.
Agentic Infrastructure
Software Engineer, Cybersecurity Researcher & Open-Source Developer building AI infrastructure, MCP servers, autonomous agent systems, and developer tools. Creator of AgentFlight and other projects focused on AI security, observability, automation, and reliable agentic systems.
FanQieAPI
FanQieAPI (https://fanqieapi.com/) is committed to providing a unified model service platform for enterprises and agents, with team-based self-deployment of open-source large models and pay-per-use access, helping enterprises and agents obtain low-cost AI computing power simply and efficiently.
Abhigya AI
OpenAI-compatible serverless inference. Agents call chat completions and related LLM endpoints on demand through one hosted API — pay for what you use, no capacity to manage.
nomad-text-api
Nomad Text API offers OpenAI-compatible text chat with ten named models: GLM 5.3 and 5.2, Nemotron 3 Ultra and Super, gpt-oss 120B and 20B, Gemma 4 31B and 26B A4B, Llama 4 Maverick, and Mistral Large 3. All ten models have passed native streaming, final token usage and canonical-label checks at the Nomad public endpoint. OpenModels forwarding has not yet been verified; model routes remain not ready. Per-model input/output token pricing is published on our website. Shared capacity: 1,000 inference attempts per rolling 24 hours; this channel is limited to 600 attempts. Concurrency 1, 2,000 total input characters, up to 16 messages and 4,096 output tokens including reasoning. No automatic model substitution. Our relay does not store prompt or response text in its application journal; external processing location, retention and training policies remain unverified. Privacy: https://inference.nomadsustaintech.com/privacy
CedarLLM
Open source models like qwen3.8-27b.
hong3yang
this is hong3yang‘s describe

