Skip to content
View CaspianG's full-sized avatar

Block or report CaspianG

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
CaspianG/README.md

CaspianG builds adaptive memory infrastructure

CaspianG

Building adaptive memory systems for software that needs context, history, and reliable recall.

WaveMind PyPI Tests Stars

WaveMind  ·  PyPI  ·  Quick Start  ·  Benchmarks  ·  Contribute

What I Build

I work on systems that make memory usable: storage, retrieval, priority, decay, feedback, corrections, and benchmarks that show whether recall actually improves over time.

My current focus is WaveMind, a local-first adaptive memory layer. Vector search finds candidates; memory state decides what deserves attention now.

WaveMind

Adaptive memory for software that remembers, forgets, and evolves.

Python API, CLI, FastAPI, SQLite/Postgres storage, ANN indexes, namespaces, TTL, feedback, graph recall, Studio, Docker, Helm, benchmarks, and production-readiness checks.

Repository  ·  PyPI  ·  Quick Start

WaveMind memory loop: store, search, memory state, recall, feedback 
pip install wavemind
wavemind quickstart
wavemind studio

Current Focus

Track Shipping direction
Memory quality Hotness, decay, corrections, TTL, feedback signals, stale-fact suppression, and graph recall.
Developer experience Python API, CLI, FastAPI server, Studio UI, imports, backups, and framework examples.
Scale path SQLite/Postgres truth stores, ANN candidate indexes, sharding, cache layers, and reproducible scale evidence.
Evidence Long-memory benchmarks, retrieval baselines, latency profiles, regression tests, and public result artifacts.

Projects

Project Snapshot Links
WaveMind Adaptive memory infrastructure for software that must preserve, prioritize, and update context over time. Docs / Benchmarks / Issues
focus-flow Minimal desktop focus timer for deep-work sessions with themes and English/Russian UI. Repository
CORECITY Browser game experiment around a living market mechanic driven by players. Repository

Stack

Python FastAPI SQLite PostgreSQL FAISS Qdrant Redis Docker Kubernetes TypeScript

Open To

Collaboration Useful contribution
Benchmarks Long-memory evaluation, stale-fact suppression, retrieval quality, latency, agent-impact tests, and scale profiles.
Integrations LangChain, LangGraph, LlamaIndex, CrewAI, AutoGen, notebooks, local apps, and production workflows.
Production feedback Real systems where memory must evolve, forget, explain, or preserve user-specific context over time.

Contact

Open an issue in WaveMind if you want to test the project, contribute an integration, add a benchmark, or discuss adaptive memory for production software.

Popular repositories Loading

  1. wavemind wavemind Public

    Local-first dynamic memory for agents and apps: SQLite source of truth, vector search candidates, hotness, decay, TTL, namespaces, and benchmarks.

    Python 5 2

  2. focus-flow focus-flow Public

    Minimal desktop focus timer for 50/10 and 90/15 deep-work sessions with session planning, light/dark themes, and English/Russian UI.

    TypeScript 1

  3. CORECITY CORECITY Public

    Онлайн-игра в формате «живой биржи», где рынок двигают не скрипты, а игроки.

    JavaScript

  4. CaspianG CaspianG Public

    Dynamic memory infrastructure, WaveMind, and local-first systems.