convez.ai
MCP ServerProdutividadeAvançado
TaeSooPark-PTS

Lattice AI

TaeSooPark-PTS/LatticeAI

Cérebro IA privado que mantém seu conhecimento durável entre modelos

Ver no GitHub
BomPluginScore · de 0 a 100

Espaço de trabalho IA local-first que preserva conhecimento, conversas e contexto de projetos no seu computador, independente do modelo usado. Integra chat web, VS Code, Telegram, MCP servers e Graph RAG com foco em privacidade e controle total sobre sua memória IA.

Instalação

$npx convez-ai TaeSooPark-PTS/LatticeAI

1 comando no terminal: detecta o formato do repositório (marketplace ou plugin avulso) e instala via Claude Code. Requer Node 18+ e o Claude Code instalado.

Métricas

Estrelas
2
Forks
0
Acompanhando
0
Saúde
100%
Último commit
há 1 dia
Versão
v13.1.0
Issues abertas
0
Contribuidores
3

Casos de uso

  • Preservar contexto de projeto ao trocar entre diferentes modelos de IA
  • Centralizar decisões, documentos e conversas em uma memória pessoal estruturada
  • Usar IA local sem perder dados em serviços em nuvem
  • Manter auditoria, backup e restauração completa do seu conhecimento IA
  • Automatizar workflows de memorização e digests de conhecimento com consentimento explícito

Veredito

Solução madura e bem-estruturada para quem precisa de memória IA persistente e privada. Excelente integração com VS Code e suporte multiplataforma, mas requer conforto com arquitetura local-first e setup inicial mais robusto. Recomendado para profissionais que lidam com múltiplos modelos e necessitam auditoria completa do conhecimento.

README

A tradução para o português deste README ainda está na fila.

Lattice AI

Lattice AI is a local-first Digital Brain that keeps your knowledge durable across any AI model.

Lattice AI는 모델이 바뀌어도 내 지식과 맥락을 보존하는 로컬 우선 AI 브레인입니다.

Your model is the voice you use today. Your Brain is the asset you keep. Lattice AI preserves conversations, documents, decisions, project context, relationships, and workflows on your computer by default. Cloud models, model downloads, update checks, and other external communication happen only after explicit consent.

It is not a ChatGPT clone, a model launcher, a graph database, or a note app. It is a private AI memory layer wrapped in a Living Brain experience.

PyPI Version npm Version VS Code Marketplace Version Open VSX Version CI Status License

Why You Need It

You need Lattice AI when:

  • you ask different AI models about the same project and lose the context each time;
  • your decisions are scattered across chats, notes, PDFs, folders, and tools;
  • you want to switch models without rebuilding memory from zero;
  • you want your AI Brain to stay on your computer by default;
  • you want backup, restore, inspect, and export paths for your Brain.

이런 사람에게 필요합니다:

  • 매번 AI를 바꿀 때마다 프로젝트 맥락을 다시 설명하는 사람
  • 문서, 대화, 결정, 파일이 여기저기 흩어져 있는 사람
  • 내 지식을 특정 AI 서비스 안에 묶어두고 싶지 않은 사람
  • 로컬에 저장되는 개인 AI 브레인을 원하는 사람

What You Can Do

  • Chat with a Brain that remembers useful context instead of treating every session as disposable.
  • Add documents, local folders, notes, screenshots, and conversations with source-aware memory.
  • See recent memories, older memories, topics, relationships, and the full knowledge graph when you want deeper structure.
  • Create consent-first Brain automation drafts for memory digests, project reviews, and follow-up suggestions before any schedule is enabled.
  • Use a recommended local model without learning model internals first.
  • Keep advanced controls, audit logs, roles, and retention in a separate Admin surface.
  • Export or back up your Brain as an encrypted .latticebrain archive.

One-Minute Flow

  1. Launch the app.
  2. Create or open a local profile.
  3. Let Lattice explain what this computer can run.
  4. Start with the recommended model, or skip and choose later.
  5. Talk to your Brain.
  6. Watch memories, topics, relationships, and graph structure emerge from real use.
  7. Back up, inspect, export, or restore the Brain when you need ownership actions.

Living Brain Flow

1. Login

Choose the owner of the Brain. The profile is not a SaaS account by default; it is the local identity for the knowledge you keep.

Login

2. Environment Analysis

See what kind of local AI experience this computer can support before choosing a model.

Environment Analysis

3. Recommended Models

Start with a short list: safest recommendation, faster model, stronger model. Advanced details stay available without overwhelming first-time users.

Recommended Models

4. Install And Load

Download and load only after consent. Lattice explains model size, local execution, and network use before work starts.

Install and Load

5. Brain Chat

Talk normally. Useful decisions and context become memory, then appear later as topics, relationships, and graph structure.

Brain Chat Home

6. Review Center

Automation results are staged for review before they become durable decisions. Snooze, unsnooze, run now, approve, and dismiss actions stay explicit.

Review Center

Brain Depths

The user travels inward from everyday memory to deeper structure:

LevelUser nameWhat the user gets
Level 1Now memoryThe living Brain presence and current conversation context
Level 2Older memoryDurable memories with source-aware recall
Level 3TopicsRecurring themes across chats and documents
Level 4RelationshipsHow decisions, people, files, and ideas connect
Level 5Full knowledge graphNodes, edges, search, and focused detail for advanced exploration

Walkthrough:

v6.1.0 Living Brain walkthrough

Screenshot index and capture notes: output/release/v6.1.0/SCREENSHOT_INDEX.md

Install

Run from Python:

pip install ltcai
LTCAI

Run from npm:

npm install -g ltcai
ltcai

Open the local app:

http://127.0.0.1:4825/app

Apple Silicon local model extras:

pip install "ltcai[local]"

Architecture At A Glance

  • Product category: local-first Digital Brain.
  • Core capability: private AI memory layer for conversations, documents, decisions, relationships, workflows, and project context.
  • UX metaphor: Living Brain.
  • Desktop shell: Tauri 2 starts a localhost sidecar.
  • Frontend: React, TypeScript, Vite, TanStack Query, Zustand, Cytoscape.js, React Flow, and generated OpenAPI types.
  • Backend: FastAPI on localhost is the UI source of truth.
  • Brain Core: independent lattice_brain package for graph, memory, context, conversations, ingestion, runtime, workflow, storage, and portability.
  • Storage: SQLite default; PostgreSQL/pgvector is optional scale mode.
  • Portability: encrypted .latticebrain archives plus backup, restore, inspect, verify, import dry-run, and confirmed restore/import flows.
  • Trust boundary: local-first by default; cloud calls, downloads, Telegram, Brain Network, Docker/Postgres setup, and update checks are opt-in.
  • Admin separation: normal Brain use stays separate from users, audit logs, policies, security events, retention, and index rebuilds.

See ARCHITECTURE.md for the current architecture.

Local Development

npm install
npm run dev

Main validation set:

npm run check:python
node scripts/run_python.mjs -m ruff check .
npm run lint
npm run typecheck
npm run test:unit
npm run test:integration
npm run test:visual
npm run desktop:tauri:check
npm run docs:check-links

See docs/DEVELOPMENT.md for developer workflow details.

Current Release Preparation

The current development target is 6.1.0 Product Hardening / Digital Brain Completion:

  • First-run now lets users create or open the local Brain even when they defer model setup, keeping the Brain durable while models stay replaceable.
  • Brain Home explains the first memory loop: save useful context, see it return in the Brain state, then protect it with backup/export.
  • Review Center keeps Pending, Snoozed, and All filters while clarifying that Run now is a preview/regenerate action, not approval.
  • lattice_brain has an AST import guard so Brain Core does not import latticeai or ltcai.
  • Model download consent has a unit guard: token presence alone does not start an external download path.
  • ltcai_cli.py keeps the root entrypoint while pure CLI runtime helpers live under latticeai.cli.runtime.
  • Tool dispatch authorization now has an injectable service boundary around the shared ToolRegistry, reducing app-factory reliance on module globals.
  • Chat agent runtime construction moved into app-factory assembly and is passed through AppContext, so the chat router no longer owns production runtime construction.
  • v6.1 documentation tracks backend hardening, frontend UX hardening, and the baseline scan for version metadata, root legacy modules, Brain Core boundary, and local-first trust gates.

Expected artifacts for 6.1.0 release must use exact filenames:

  • dist/ltcai-6.1.0-py3-none-any.whl
  • dist/ltcai-6.1.0.tar.gz
  • ltcai-6.1.0.tgz
  • dist/ltcai-6.1.0.vsix
  • src-tauri/target/release/bundle/dmg/Lattice AI_6.1.0_aarch64.dmg

Do not upload dist/*. Package registry publishing remains owner-run.

Known Limitations

  • External package registries are owner-published and can lag behind GitHub.
  • PostgreSQL/pgvector is optional scale mode. SQLite is the default local Brain.
  • Docker, model downloads, cloud model calls, Telegram, Brain Network, and update checks require explicit user action.
  • Conversation does not fabricate answers when no model is loaded.
  • Agent/workflow simulation without a loaded LLM is deterministic and does not call a model; it is labeled as LLM-free/model-free rather than presented as autonomous model success.

Release History

VersionTheme
6.0.0Product Reset / Review Center Completion: Snoozed filter, Unsnooze, OpenAPI-derived Review typing, Review feature extraction, v6 docs and scorecard
5.6.0Brain Automation Review Center: workspace-scoped automation review inbox, source-aware provenance, guarded approve/dismiss/snooze/run_now actions, and Act Review tab
5.5.0Release Coordination: synchronized package/runtime/static versions and release docs for the 5.5.0 line while preserving v5.4.0 Brain Automation Scheduler behavior
5.4.0Brain Automation Scheduler: consent-first recipe drafts (Daily/Weekly/Follow-up), TriggerService with dedup/LATTICE_TZ/degraded, runtime graph cleanup, E2E scenarios
5.3.0Product Clarity and Runtime Cleanup: user-first README/onboarding, unified Digital Brain identity, legacy compatibility map, and app factory runtime seams
5.2.0User-Focused Model Transformation: structured model capability registry, HF verification transparency, model recommendation UX, and workspace-scoped marketplace state
5.1.0Product Trust & Clarity Release: clarifies the private AI memory-layer promise, hardens CSP/secret/auto-read/download gates, adds trust/privacy docs, and refreshes evidence
5.0.0Multilingual Brain Foundation Release: adds persisted Korean/English language choice across first-run onboarding, Brain home, graph exploration, and Admin Console
4.7.2Intuitive Brain UX Release: safer login, one-click recommended setup, direct Brain views, memory-save feedback, and exact artifacts
4.7.0Admin Separation Release: adds the separate Admin Console for users/logs/security/Brain operations
4.6.1Living Brain Release Refresh: publishable version bump after v4.6.0 PyPI immutability
4.6.0Living Brain Experience: made Brain plus conversation the home product
4.5.1Product Reimagining RC: reset shell, navigation, onboarding, and visual system
4.5.0Product Experience Recovery RC: restored first-run setup, workspace/model onboarding, and graph discoverability

Documentation

License

MIT. See LICENSE.

Ver o original no GitHub

Bom para

Relacionados