# Reminix

## Docs

- [Chat with an agent](https://reminix.com/docs/api-reference/agents/chat-with-an-agent.md): Send a chat message to an agent and receive a response.
- [Get an agent](https://reminix.com/docs/api-reference/agents/get-an-agent.md): Get details of a specific agent by name.
- [Invoke an agent](https://reminix.com/docs/api-reference/agents/invoke-an-agent.md): Invoke an agent with the provided input.
- [List agents](https://reminix.com/docs/api-reference/agents/list-agents.md): List all agents in the project with optional filtering by stack and status.
- [Start a workflow run](https://reminix.com/docs/api-reference/agents/start-a-workflow-run.md): Start a new workflow run for a workflow agent.
- [Create a client token](https://reminix.com/docs/api-reference/client-tokens/create-a-client-token.md): Create a short-lived client token for browser SDK use.
- [Revoke a client token](https://reminix.com/docs/api-reference/client-tokens/revoke-a-client-token.md): Revoke a client token immediately.
- [Delete a conversation](https://reminix.com/docs/api-reference/conversations/delete-a-conversation.md): Delete a conversation and all its messages. This action cannot be undone.
- [Get a conversation](https://reminix.com/docs/api-reference/conversations/get-a-conversation.md): Get details of a specific conversation including all messages.
- [List conversations](https://reminix.com/docs/api-reference/conversations/list-conversations.md): List conversations for the project with optional filtering.
- [Introduction](https://reminix.com/docs/api-reference/introduction.md): REST APIs for agents, Model Context Protocol (MCP) servers for tools. Invoke, chat, stream, and manage workflows.
- [Delete an OAuth connection](https://reminix.com/docs/api-reference/oauth-connections/delete-an-oauth-connection.md): Delete an OAuth connection by name. This revokes access and removes the stored tokens. This action cannot be undone.
- [Get a fresh access token for a connection](https://reminix.com/docs/api-reference/oauth-connections/get-a-fresh-access-token-for-a-connection.md): Get a fresh access token for an OAuth connection. The token is automatically refreshed if expired.
- [Get OAuth connection by name](https://reminix.com/docs/api-reference/oauth-connections/get-oauth-connection-by-name.md): Get details of a specific OAuth connection by its name (slug), including provider info, granted scopes, and status.
- [List OAuth connections](https://reminix.com/docs/api-reference/oauth-connections/list-oauth-connections.md): List OAuth connections for the project with optional filtering by provider.
- [Get current project](https://reminix.com/docs/api-reference/projects/get-current-project.md): Returns the project associated with the API key or X-Project header.
- [Delete a workflow run](https://reminix.com/docs/api-reference/workflow-runs/delete-a-workflow-run.md): Delete a workflow run. This action cannot be undone.
- [Get a workflow run](https://reminix.com/docs/api-reference/workflow-runs/get-a-workflow-run.md): Get details of a specific workflow run including its full state.
- [List workflow runs](https://reminix.com/docs/api-reference/workflow-runs/list-workflow-runs.md): List workflow runs for the project with optional filtering.
- [Resume a paused workflow run](https://reminix.com/docs/api-reference/workflow-runs/resume-a-paused-workflow-run.md): Resume a paused workflow run by providing the required action response.
- [Coding Assistants](https://reminix.com/docs/cli/ai-agents.md): Use the Reminix CLI as a tool inside Claude Code, Cursor, Windsurf, and other coding assistants.
- [CI/CD](https://reminix.com/docs/cli/ci-cd.md): Run the Reminix CLI from GitHub Actions and other CI/CD systems — invoke agents, mint client tokens, and automate deploys.
- [CLI Reference](https://reminix.com/docs/cli/index.md): Install the CLI, authenticate, deploy projects, and manage agents, secrets, workflows, and execution logs — all from your terminal.
- [Scripting](https://reminix.com/docs/cli/scripting.md): Invoke agents, call tools, and pipe --output json through jq and shell scripts to automate Reminix from any Unix environment.
- [Agents](https://reminix.com/docs/concepts/agents.md): Define handlers in TypeScript or Python, pick from prompt, chat, task, thread, or workflow types, and serve via REST API.
- [Authentication](https://reminix.com/docs/concepts/authentication.md): Three Reminix token types — Project API Keys, Personal Access Tokens, and Client Tokens — and when to use each.
- [Configuration & Secrets](https://reminix.com/docs/concepts/configuration.md): Set project secrets in the dashboard or via reminix secret set, then read them from your handler as standard environment variables.
- [OAuth & Connections](https://reminix.com/docs/concepts/connections.md): Managed OAuth for 20+ services like Google, Slack, GitHub, and Notion. Your code gets a valid token; Reminix handles the flow and refresh.
- [Conversations](https://reminix.com/docs/concepts/conversations.md): Multi-turn conversations with persistent message history, conversation IDs, and identity scoping for multi-tenant apps.
- [Errors](https://reminix.com/docs/concepts/errors.md): Reminix error response format, HTTP status codes, and the execution URL that links every failure to full logs in the dashboard.
- [Streaming](https://reminix.com/docs/concepts/streaming.md): Stream agent responses token-by-token over Server-Sent Events. Works with tasks, conversations, and workflows in both SDKs.
- [Tasks](https://reminix.com/docs/concepts/tasks.md): The simplest interaction pattern — send input, get output, with no conversation state or persistence between calls.
- [Tools & MCP](https://reminix.com/docs/concepts/tools.md): Define functions as tools and serve them via the Model Context Protocol so Claude, Cursor, and Windsurf can call them automatically.
- [Troubleshooting](https://reminix.com/docs/concepts/troubleshooting.md): Common deploy failures, runtime errors, MCP connection issues, and streaming problems — with the symptom, cause, and fix for each.
- [Workflows](https://reminix.com/docs/concepts/workflows.md): Multi-step agent processes that can pause, resume, and request human input — for approval flows, batch jobs, and long-running pipelines.
- [Deploy Button](https://reminix.com/docs/deploy-button.md): Add a clickable 'Deploy to Reminix' button to your repo's README so visitors can fork and deploy your project in one click.
- [Deploy from GitHub](https://reminix.com/docs/deploy-from-github.md): Connect a GitHub repository to Reminix, deploy on every push, monitor build logs, and add a one-click deploy badge to your README.
- [What is Reminix?](https://reminix.com/docs/index.md): You built the agent — now ship it. Deploy agents as REST APIs and tools as MCP servers with streaming, auth, and monitoring. TypeScript and Python.
- [Creating Agents](https://reminix.com/docs/python/agents/creating-agents.md): Define agents in Python with the @agent decorator, derive schemas from type hints, and expose them via serve().
- [Creating Tools](https://reminix.com/docs/python/agents/creating-tools.md): Define MCP-compatible tools in Python with the @tool decorator and type hints — automatically discoverable by any MCP client.
- [Deploying](https://reminix.com/docs/python/agents/deploying.md): Package and deploy Python agents and tools, with notes on entrypoints, dependencies, and runtime env vars.
- [Chat](https://reminix.com/docs/python/chat.md): Send chat messages to your agents from Python and keep the conversation going via Reminix-managed conversation IDs.
- [Error Handling](https://reminix.com/docs/python/error-handling.md): Catch typed exceptions like APIError, AuthenticationError, and RateLimitError, and add retry logic for transient failures.
- [Anthropic](https://reminix.com/docs/python/frameworks/anthropic.md): Wrap an AsyncAnthropic client as a Reminix chat or task agent with Claude models, streaming, and structured output.
- [Google AI](https://reminix.com/docs/python/frameworks/google.md): Wrap a genai.Client as a Reminix chat or task agent with Gemini models, streaming, structured output, and tool calls.
- [LangChain](https://reminix.com/docs/python/frameworks/langchain.md): Serve LangChain Runnables, ChatModels, and compiled LangGraph agents as Reminix chat, task, or thread agents with streaming.
- [OpenAI](https://reminix.com/docs/python/frameworks/openai.md): Wrap an AsyncOpenAI client as a Reminix chat or task agent with streaming, structured output, and the official OpenAI Python SDK.
- [Python SDK](https://reminix.com/docs/python/index.md): Install the Reminix Python SDK, configure your API key, and call agents and tools from sync or async Python code.
- [Invoke](https://reminix.com/docs/python/invoke.md): Call task agents from Python with one request and one response. Supports streaming, idempotency keys, and async usage.
- [Quickstart](https://reminix.com/docs/quickstart.md): Create a project, deploy your first agent from a template, and call it from your app — under five minutes, no local setup required.
- [Creating Agents](https://reminix.com/docs/typescript/agents/creating-agents.md): Define agents in TypeScript with the agent() factory, pick a type, attach a handler, and expose them via serve().
- [Creating Tools](https://reminix.com/docs/typescript/agents/creating-tools.md): Define MCP-compatible tools in TypeScript with the tool() factory and Zod schemas — automatically discoverable by any MCP client.
- [Deploying](https://reminix.com/docs/typescript/agents/deploying.md): Package and deploy TypeScript agents and tools, with notes on entrypoints, dependencies, and runtime env vars.
- [Chat](https://reminix.com/docs/typescript/chat.md): Send chat messages to your agents from TypeScript and keep the conversation going via Reminix-managed conversation IDs.
- [Error Handling](https://reminix.com/docs/typescript/error-handling.md): Catch typed error classes like APIError, AuthenticationError, and RateLimitError, and add retry logic for transient failures.
- [Anthropic](https://reminix.com/docs/typescript/frameworks/anthropic.md): Wrap any Anthropic client as a Reminix chat or task agent with Claude models, streaming, and structured output.
- [Google AI](https://reminix.com/docs/typescript/frameworks/google.md): Wrap a GoogleGenAI client as a Reminix chat or task agent with Gemini models, streaming, and structured output.
- [LangChain](https://reminix.com/docs/typescript/frameworks/langchain.md): Serve LangChain runnables, ChatModels, and compiled LangGraph agents as Reminix chat, task, or thread agents with streaming.
- [OpenAI](https://reminix.com/docs/typescript/frameworks/openai.md): Wrap any OpenAI client as a Reminix chat or task agent with streaming, structured output, and the official OpenAI SDK.
- [Vercel AI SDK](https://reminix.com/docs/typescript/frameworks/vercel-ai.md): Wrap Vercel AI SDK models, ToolLoopAgents, and chat flows as Reminix agents with streaming, structured output, and tool calls.
- [TypeScript SDK](https://reminix.com/docs/typescript/index.md): Install the Reminix TypeScript SDK, configure your API key, and call agents and tools from your TypeScript or Node.js code.
- [Invoke](https://reminix.com/docs/typescript/invoke.md): Call task agents from TypeScript with one request and one response. Supports streaming, idempotency keys, and execution context.

## OpenAPI Specs

- [openapi.documented](https://app.stainless.com/api/spec/documented/reminix/openapi.documented.yml)

## Optional

- [Dashboard](https://reminix.com/dashboard)
- [GitHub](https://github.com/reminix-ai)