Agentic Integrations
Built for Agents
Seminara is designed to be commanded by agents, not just humans. Whether you use a chat assistant, a local CLI, or a desktop AI tool, Seminara exposes the same underlying capabilities through a consistent, agent-friendly interface.
Choose your integration method below.
Approve a Device Connection
If your CLI tool or AI agent has displayed an activation code, visit the link below and enter the code to authorize it. Your access token is generated instantly on approval.
Activate DeviceFrequently asked questions.
Everything you need to know about this resource.
An authentication standard (RFC 8628) that lets external agents or CLI tools prompt you with a short code (e.g. AAAA-BBBB) to link your account securely, without ever seeing your password.
No. The CLI, MCP server, and any agent that supports the Device Flow will request authorization automatically. You only need to visit /activate and enter the code when prompted.
Very secure. The MCP server runs locally on your machine and only communicates with the Seminara API over HTTPS. Your access token is stored in ~/.config/seminara/auth.json and never transmitted elsewhere.
Any agent that can make HTTP requests (Claude, ChatGPT, Gemini, OpenCode, Cursor, Codex). The Skill prompt is designed to work with any modern LLM regardless of how it implements tool calling.
Ready to Integrate?
Generate your access token from the Settings page and give your agents the power to host.