From 010a791ffe6408245b64cca7e78d37ff02ce21cb Mon Sep 17 00:00:00 2001 From: "github-actions[bot]" Date: Fri, 5 Jun 2026 14:50:07 +0000 Subject: [PATCH] docs: update for pipecat-flows PR #281 Update documentation to reflect changes from pipecat-flows PR #281: - Add append_text_to_context parameter to ActionConfig - Document new parameter for tts_say and end_conversation actions - Document initial node context strategy behavior change These changes align with the new behavior where: 1. Built-in TTS actions now support controlling whether spoken text is appended to the LLM context (defaults to True) 2. Initial flow nodes now follow their context strategy instead of always replacing the context --- api-reference/pipecat-flows/types.mdx | 13 +++++++---- pipecat-flows/guides/actions.mdx | 24 +++++++++++++++++++++ pipecat-flows/guides/context-strategies.mdx | 7 ++++++ 3 files changed, 40 insertions(+), 4 deletions(-) diff --git a/api-reference/pipecat-flows/types.mdx b/api-reference/pipecat-flows/types.mdx index 045ad3d3..f93b12f6 100644 --- a/api-reference/pipecat-flows/types.mdx +++ b/api-reference/pipecat-flows/types.mdx @@ -186,8 +186,13 @@ TypedDict for configuring actions that execute during node transitions. - Text content used by `tts_say` and optionally by `end_conversation` (as a - goodbye message). + Text to speak for the `tts_say` action, or the optional goodbye message for + the `end_conversation` action. + + + + For the built-in TTS actions (`tts_say` and `end_conversation`), whether the + spoken text is appended to the LLM context. Defaults to `True`. @@ -199,8 +204,8 @@ TypedDict for configuring actions that execute during node transitions. | Type | Description | Required Fields | | ------------------ | ----------------------------------------------------------- | ----------------- | -| `tts_say` | Speak text using the pipeline's TTS service | `text` | -| `end_conversation` | End the conversation, optionally speaking a goodbye message | `text` (optional) | +| `tts_say` | Speak text using the pipeline's TTS service. The spoken text is appended to the LLM context by default; set `append_text_to_context=False` to opt out. | `text` | +| `end_conversation` | End the conversation, optionally speaking a goodbye message. If `text` is provided, it is appended to the LLM context by default; set `append_text_to_context=False` to opt out. | `text` (optional) | | `function` | Execute a function inline in the pipeline | `handler` | ### Example diff --git a/pipecat-flows/guides/actions.mdx b/pipecat-flows/guides/actions.mdx index 789d5a60..cf7dd3ed 100644 --- a/pipecat-flows/guides/actions.mdx +++ b/pipecat-flows/guides/actions.mdx @@ -27,6 +27,18 @@ Speak a phrase immediately (useful for "please wait" messages): ] ``` +The spoken text is appended to the LLM context by default. To prevent the text from being added to the context, set `append_text_to_context=False`: + +```python +"pre_actions": [ + { + "type": "tts_say", + "text": "Processing...", + "append_text_to_context": False + } +] +``` + ### end_conversation Gracefully terminate the conversation: @@ -40,6 +52,18 @@ Gracefully terminate the conversation: ] ``` +The goodbye text is appended to the LLM context by default. To prevent the text from being added to the context, set `append_text_to_context=False`: + +```python +"post_actions": [ + { + "type": "end_conversation", + "text": "Goodbye!", + "append_text_to_context": False + } +] +``` + ### function Execute a custom function at the specified timing: diff --git a/pipecat-flows/guides/context-strategies.mdx b/pipecat-flows/guides/context-strategies.mdx index 9eb3bbda..71f70a2a 100644 --- a/pipecat-flows/guides/context-strategies.mdx +++ b/pipecat-flows/guides/context-strategies.mdx @@ -20,6 +20,13 @@ Flows provides three built-in ways to manage conversation context as you move be pre-action. + + **Initial node behavior:** The initial flow node follows its context strategy (default `APPEND`) + instead of always replacing the context. With the default strategy, context written before the + node's first inference — for example by a `tts_say` pre-action — is now preserved rather than + discarded. Set `context_strategy=RESET` on the initial node for the previous clean-slate behavior. + + ## When to Use Each Strategy - Use **APPEND** when full conversation history is important for context