Ana Sayfa / Kod ve Scriptler / Voice Memory SDK for Python and Node

Voice Memory SDK for Python and Node

Voice Memory, yerel depolama ve Python ile Node SDK'ları üzerinden ses ajanlarına arayan kişiye özel özetler ekler.

  • Yayın tarihi11 Kasım 2026
  • Paket içeriği.py, .json, .md, package.json, JavaScript .js
  • Son güncelleme11 Kasım 2026
  • Canlı demoVar
Voice Memory SDK for Python and Node

Overview

Voice Memory sits between your voice framework and a single local JSON file. When a call ends, it reads the transcript and extracts what is worth keeping — preferences, open threads, facts — and scrubs personal data out before anything is stored. On the next call from the same person, it builds a short brief you drop into the system prompt, and it can plan the call flow to skip steps the memory already covers.

Voice platforms charge extra for per-contact memory, and the free stacks (Pipecat, LiveKit) ship nothing like it — every call starts from zero and regulars repeat themselves. Voice Memory is the missing layer, and it is deliberately small: one store file you can open and read, two SDKs (Node and Python) that read and write the same file, and adapters for Pipecat and LiveKit.

What it does:

- recordCall: extract preferences, open threads, and facts from a finished call's transcript

- buildInjection: a short per-caller brief, ready to paste into the next call's system prompt

- planFlow: skip scripted steps the memory already answers

- Personal data scrubbed before storage — the redaction pass runs on the way in, not on request

- One JSON store file shared by both SDKs: Node (src/) and Python (python-sdk/) are fully interoperable

- Adapters for Pipecat and LiveKit included, plus a plain base adapter to copy

- Demo with two fictional calls from the same caller, so you can watch the memory build

- Legacy callrecall-ai stores auto-upgrade on load; CallRecallMemory still imports as an alias

What's inside the download:

- Voice Memory source: Node SDK (src/), Python SDK (python-sdk/), Pipecat frame processor (pipecat/)

- README.md — what it is and how to run it

- INSTALL.md — step-by-step setup for both SDKs

- SPEC.md — the store schema and extraction rules, in full

- demo/ — two mock calls, a runnable demo script, and the store it produces

- examples/ — Pipecat and LiveKit wiring examples

- Test suites for both SDKs (smoke + end-to-end)

- LICENSE — MIT license

How you run it: Run `node demo/demo.cjs` first: call one ends, the memory is written; call two opens with the brief already in the prompt. Then wire recordCall / buildInjection into your own Pipecat or LiveKit pipeline using the examples.

Open source, stated plainly: Voice Memory is my own code, and the same code is free on GitHub under the MIT license. This package is the supported download — the tested release zipped with its docs, examples, and test suite, and email support from me if something in the package is broken. If you'd rather clone the repo, that door is open. I'd rather you knew before you paid.

You pay once — there is no subscription and no account to create.