maximem-synap-pipecat 0.2.0

Synap memory integration for Pipecat — voice-pipeline context injection + conversation recording

synap-pipecat

Synap integration for Pipecat — preload long-term memory into voice pipelines and record every turn back to Synap.

Install

pip install synap-pipecat

Requires pipecat-ai>=0.0.80, maximem-synap>=0.2.0.

Quickstart

from pipecat.pipeline.pipeline import Pipeline
from pipecat.processors.aggregators.llm_context import LLMContext
from pipecat.processors.aggregators.llm_response_universal import (
    LLMContextAggregatorPair,
    LLMUserAggregator,
    LLMAssistantAggregator,
)
from maximem_synap import MaximemSynapSDK
from synap_pipecat import SynapMemoryProcessor, SynapRecorder

sdk = MaximemSynapSDK(api_key="sk-...")
context = LLMContext(messages=[{"role": "system", "content": "You are a helpful assistant."}])

user_agg = LLMUserAggregator(context=context)
asst_agg = LLMAssistantAggregator(context=context)

pipeline = Pipeline([
    transport.input(),
    stt,
    SynapMemoryProcessor(sdk, user_id="alice", customer_id="acme"),
    user_agg,
    llm,
    tts,
    transport.output(),
    asst_agg,
    SynapRecorder(sdk, user_id="alice", customer_id="acme"),
])

Scope

• SynapMemoryProcessor — sits BEFORE the user context aggregator. On each TranscriptionFrame, fetches relevant context from Synap via sdk.fetch() and appends it as a system message to the shared LLMContext. The transcription frame itself flows through unchanged so the user aggregator can record it as the user turn.
• SynapRecorder — sits AFTER the assistant context aggregator. Buffers the latest user transcription and the streamed assistant response; on LLMFullResponseEndFrame, calls sdk.conversation.record_message() for both turns.

Error policy

• Reads (SynapMemoryProcessor) degrade gracefully — SDK failures log at ERROR and the frame continues downstream with no Synap context injected. A Synap blip must never break a live voice call.
• Writes (SynapRecorder) surface SynapIntegrationError via wrap_sdk_errors_async. The processor logs the failure and pushes an ErrorFrame upstream; the pipeline itself stays alive (errors propagate as frames, not exceptions).
• Non-matching frames pass through untouched in both processors — standard Pipecat push_frame contract.

Constructor args (both processors)

• sdk: a configured MaximemSynapSDK
• user_id: required — Synap memory is user-scoped
• customer_id: optional customer/org scope; empty string means customer-less
• conversation_id: optional explicit conversation id (auto-generated per process lifetime when absent, shared across both processors if you pass the same instance)