# Autoplay SDK

> Powering real-time and context Aware copilots

## Docs

- [Changelog](https://developers.autoplay.ai/changelog.md): Version history and release notes for the autoplay-sdk Python package.
- [Botpress](https://developers.autoplay.ai/integrations/botpress.md): Dedicated Botpress integration helpers are coming soon. Stream events to your connector and chatbot today.
- [Dify](https://developers.autoplay.ai/integrations/diffy.md): Dedicated Diffy integration helpers are coming soon. Stream events to your connector and chatbot today.
- [Help Scout](https://developers.autoplay.ai/integrations/help-scout.md): Dedicated Help Scout integration helpers are coming soon. Stream events to your connector and chatbot today.
- [HubSpot Chat](https://developers.autoplay.ai/integrations/hubspot-chat.md): Dedicated HubSpot Chat integration helpers are coming soon. Stream events to your connector and chatbot today.
- [Intercom](https://developers.autoplay.ai/integrations/intercom.md): Intercom-specific SDK helpers in autoplay_sdk.integrations.intercom and how they map to the event connector.
- [Zendesk](https://developers.autoplay.ai/integrations/zendesk.md): Dedicated Zendesk integration helpers are coming soon. Stream events to your connector and chatbot today.
- [🚀 Quickstart](https://developers.autoplay.ai/quickstart.md): Stream real-time user events into your AI copilots and agents in a couple lines of code.
- [Step 1 — Connect real-time events](https://developers.autoplay.ai/recipes/ada/step-1-connect-real-time-events.md): Consume Autoplay's real-time event stream on your server and inject live user activity into Ada's AI Agent via metaFields.
- [Step 2 — Define proactive triggers](https://developers.autoplay.ai/recipes/ada/step-2-define-proactive-triggers.md): Proactively message users in Ada based on what they're doing in your product — config-driven, no code changes required.
- [Step 1 — Connect real-time events](https://developers.autoplay.ai/recipes/botpress/step-1-connect-real-time-events.md): Stream live user actions into Botpress tables and wire an Autonomous Agent to answer with real-time context.
- [Step 2 — Define proactive triggers](https://developers.autoplay.ai/recipes/botpress/step-2-define-proactive-triggers.md): Proactively message users in Botpress based on what they're doing in your product — coming soon.
- [Chameleon — How to setup](https://developers.autoplay.ai/recipes/chameleon/how-to-setup.md): Trigger a Chameleon tour from the Autoplay event stream.
- [Step 1 — Connect real-time events](https://developers.autoplay.ai/recipes/crisp-ai/step-1-connect-real-time-events.md): Set up the webhook server and event listener so Crisp replies with real-time context.
- [Step 2 — Define proactive triggers](https://developers.autoplay.ai/recipes/crisp-ai/step-2-define-proactive-triggers.md): Proactively message users in Crisp based on what they're doing in your product — coming soon.
- [Datadog — How to setup](https://developers.autoplay.ai/recipes/datadog/how-to-setup.md): Learn how to set up live user activity from Datadog to feed as context to your chatbot using the Autoplay SDK.
- [Step 1 — Connect real-time events](https://developers.autoplay.ai/recipes/dify-tutorial/step-1-connect-real-time-events.md): Set up the event-stream server, wire the Dify External Knowledge API, and tune background summarization.
- [Step 2 — Define proactive triggers](https://developers.autoplay.ai/recipes/dify-tutorial/step-2-define-proactive-triggers.md): Poll for active triggers, render a help toast, drive the FSM, and launch visual guidance alongside the Dify chat.
- [FullStory — How to setup](https://developers.autoplay.ai/recipes/fullstory/how-to-setup.md): Learn how to set up live user activity from FullStory to feed as context to your chatbot using the Autoplay SDK.
- [Step 1 — Connect real-time events](https://developers.autoplay.ai/recipes/intercom-tutorial/step-1-connect-real-time-events.md): Register Intercom webhooks, consume the Autoplay stream, and post internal notes into conversations.
- [Step 2 — Define proactive triggers](https://developers.autoplay.ai/recipes/intercom-tutorial/step-2-define-proactive-triggers.md): Proactively message users in Intercom based on what they're doing in your product — config-driven, no code changes required.
- [Step 1 — Connect real-time events](https://developers.autoplay.ai/recipes/landbot/step-1-connect-real-time-events.md): Set up the Landbot workflow, wire the backend server and listener, and embed the chatbot in your frontend app.
- [Step 2 — Define proactive triggers](https://developers.autoplay.ai/recipes/landbot/step-2-define-proactive-triggers.md): Proactively message users in Landbot based on what they're doing in your product — coming soon.
- [Pendo — How to setup](https://developers.autoplay.ai/recipes/pendo/how-to-setup.md): Trigger a Pendo guide from the Autoplay event stream.
- [PostHog — How to setup](https://developers.autoplay.ai/recipes/posthog/how-to-setup.md): Learn how to set up live user activity from PostHog to feed as context to your chatbot using the Autoplay SDK.
- [Step 1 — Connect real-time events](https://developers.autoplay.ai/recipes/rasa/step-1-connect-real-time-events.md): Stream live user actions into a Rasa-aware bridge using the Autoplay SDK, expose them to Rasa over HTTP, and wire the chat widget.
- [Step 2 — Define proactive triggers](https://developers.autoplay.ai/recipes/rasa/step-2-define-proactive-triggers.md): Proactively message users in Rasa based on what they're doing in your product, using agent_state_v2 and the SDK's proactive trigger registry.
- [Step 1 — Connect real-time events](https://developers.autoplay.ai/recipes/tidio/step-1-connect-real-time-events.md): Set up the Tidio Lyro AI Agent, wire the backend server and listener, and embed the chat widget in your frontend app.
- [Step 2 — Define proactive triggers](https://developers.autoplay.ai/recipes/tidio/step-2-define-proactive-triggers.md): Proactively message users in Tidio based on what they're doing in your product — coming soon.
- [Told — How to setup](https://developers.autoplay.ai/recipes/told/how-to-setup.md): Trigger a Told tour from the Autoplay event stream.
- [How to trigger a User Tour](https://developers.autoplay.ai/recipes/user-tour/overview.md): Learn how to trigger any user tour provider from your backend using the Autoplay event stream.
- [Userflow — How to setup](https://developers.autoplay.ai/recipes/userflow/how-to-setup.md): Trigger a Userflow flow from the Autoplay event stream.
- [UserGuiding — How to setup](https://developers.autoplay.ai/recipes/userguiding/how-to-setup.md): Trigger a UserGuiding flow from the Autoplay event stream.
- [Userpilot — How to setup](https://developers.autoplay.ai/recipes/userpilot/how-to-setup.md): Trigger a Userpilot flow from the Autoplay event stream.
- [Usertour — How to setup](https://developers.autoplay.ai/recipes/usertour/how-to-setup.md): Trigger a Usertour flow from the Autoplay event stream.
- [AgentContextWriter](https://developers.autoplay.ai/sdk/agent-context.md): Push real-time event context to any agent destination and keep the context window bounded with LLM-compressed summaries.
- [Agent session states](https://developers.autoplay.ai/sdk/agent-states.md): Two FSM models for reactive chat, proactive offers, guidance execution, and conservative backoff — JSON snapshots, task progress, and telemetry-friendly helpers.
- [AsyncConnectorClient](https://developers.autoplay.ai/sdk/async-client.md): Async SSE client for asyncio pipelines. Callbacks are async def coroutines.
- [build_copilot_app(...)](https://developers.autoplay.ai/sdk/build-copilot-app.md): FastAPI factory for a minimal user-keyed copilot bridge with health, context, reply, and admin reset endpoints.
- [Chatbot context assembly](https://developers.autoplay.ai/sdk/chatbot-context-assembly.md): Combine user queries, real-time product events, conversation history, and an optional knowledge base into one LLM-ready context with assemble_rag_chat_context (autoplay_sdk.rag_query).
- [BaseChatbotWriter](https://developers.autoplay.ai/sdk/chatbot-writer.md): Base class for delivering session events to any chatbot platform — handles pre-link buffering, at-link flush, and post-link debouncing so you only implement the API call.
- [compose_chat_pipeline(...)](https://developers.autoplay.ai/sdk/compose-chat-pipeline.md): Compose chat ingestion primitives in the safe default order so action writes, summarization, and callback fan-out stay consistent.
- [EventBuffer](https://developers.autoplay.ai/sdk/event-buffer.md): Pull-based event access — collect real-time events and read them whenever you need.
- [Knowledge base](https://developers.autoplay.ai/sdk/knowledge-base.md): Query your product's golden paths from Autoplay's vector database to give your chatbot structured adoption context.
- [Logging](https://developers.autoplay.ai/sdk/logging.md): Structured logging conventions when building on autoplay-sdk — module loggers, exception tracebacks, and safe extra fields.
- [Migration 0.7.4](https://developers.autoplay.ai/sdk/migration-0.7.4.md): Update deprecated autoplay_sdk import paths before 1.0.0.
- [Overview](https://developers.autoplay.ai/sdk/overview.md): The future of customer support agents isn't reactive. It's Autoplay.
- [Payload schema](https://developers.autoplay.ai/sdk/payload-schema.md): Full JSON wire format for actions and summary events from the SSE stream.
- [Step 1 — Connect real-time actions](https://developers.autoplay.ai/sdk/proactive-copilot/step-1-connect-real-time-actions.md): Ingest structured ActionsPayload streams so your copilot sees what the user is doing right now.
- [Step 2 — Add actions to LLM context](https://developers.autoplay.ai/sdk/proactive-copilot/step-2-add-actions-to-llm-context.md): Merge real-time product events, the user's question, conversation history, and optional KB retrieval into one LLM-ready context block.
- [Step 3 — Define proactive triggers](https://developers.autoplay.ai/sdk/proactive-copilot/step-3-define-proactive-triggers.md): Decide when to surface proactive help, then deliver it as chat messages with optional reply options (SDK) or as on-screen visual guidance (BYO) — with registry and FSM gating.
- [Step 4 — Connect relevant visual guidance](https://developers.autoplay.ai/sdk/proactive-copilot/step-4-enhance-with-user-memory.md): Route qualifying proactive outcomes into relevant on-screen tours so guidance appears at the right time without interrupting active reactive chat.
- [Step 5 — Enrich with user memory](https://developers.autoplay.ai/sdk/proactive-copilot/step-5-enrich-with-user-memory.md): Define the workflows your product recognises so Atlas and memory can attach workflow completion rates and mastery; then filter suggestions with cross-session user memory.
- [Proactive triggers](https://developers.autoplay.ai/sdk/proactive-triggers.md): Detect the right moment to surface proactive assistance — without coupling to any chat vendor or UI layer.
- [Authoring proactive triggers](https://developers.autoplay.ai/sdk/proactive-triggers-authoring.md): Build your first proactive trigger from scratch — context, predicates, registry, timings, and delivery.
- [Built-in proactive triggers](https://developers.autoplay.ai/sdk/proactive-triggers-builtins.md): Complete reference for every trigger shipped in the SDK catalog — when it fires, what it detects, how to tune it, and how to enable it from JSON config.
- [RAG pipeline](https://developers.autoplay.ai/sdk/rag-example.md): Embed real-time session events into a vector store for retrieval-augmented generation.
- [RagPipeline](https://developers.autoplay.ai/sdk/rag-pipeline.md): Plug-and-play boilerplate that wires real-time events to any embedding model and any vector store.
- [SessionSummarizer](https://developers.autoplay.ai/sdk/summarizer.md): Client-side context-window management — accumulate actions per session and summarise them with your own LLM when a threshold is reached.
- [ConnectorClient](https://developers.autoplay.ai/sdk/sync-client.md): Sync SSE client. Callbacks run on a dedicated worker thread — blocking I/O is safe.
- [Typed payloads](https://developers.autoplay.ai/sdk/typed-payloads.md): ActionsPayload, SummaryPayload, and SlimAction — the typed models your callbacks receive.
- [User Memory](https://developers.autoplay.ai/sdk/user-memory.md): Historical per-user context so your copilot skips mastered flows and focuses on real gaps.
- [UserSessionIndex](https://developers.autoplay.ai/sdk/user-session-index.md): User-keyed session stitching for mapping one user to recent product sessions and reading cross-session activity safely.
- [WebhookReceiver](https://developers.autoplay.ai/sdk/webhook-receiver.md): Typed push webhook receiver — HMAC verification and typed payload parsing for push-mode integrations.

## OpenAPI Specs

- [openapi](https://developers.autoplay.ai/api-reference/openapi.json)