000OPEN SOURCE

We ship the stack in public.

Five open-source projects from Neura Parse. Permissive licences, no telemetry, built to be forked and shipped. Some are research companions; others are the daily tools we run on ourselves.

Developer toolsProductivityQuantum research
github.com/neuraparseBrowse projects ↓

Public projects

GitHub stars · verified 12 Jul 2026

Permissive licences

Telemetry

Software developer working with code across a laptop and monitors in a professional workspace

The open-source portfolio now carries a professional operating view across developer tools, productivity software, and research libraries with clear workflow and governance signals.

5 public projects
MIT / Apache-2.0
No telemetry
Live model
1

Trigger

188 blocks

2

Agent

300+ apps

3

Approval

API / chat / embed

4

Surface

188 blocks

Signal

188 blocks

Signal

300+ apps

Signal

API / chat / embed

001Why we open-source

We open the layers where customers benefit from inspectability, and we keep proprietary the layers where customers pay for hard-won engineering.

01

Local-first software for engineers who already have an editor and a terminal. Designed to plug into existing workflows, not to replace them.

02

Self-hosted alternatives to closed SaaS. Bring your own database, your own keys, your own infrastructure — keep your data where it belongs.

03

Reference libraries that accompany our published research. Apache-2.0, designed for academic citation and partner integration, not for vendor lock-in.

002Compare

Sort by stars, language, status, or filter by category. The detailed cards below dive into capabilities and the install path.

5 / 5

Self-hosted PM with AI-agent governance — v0.7.11, 30 locales, Docker, AGENTOWNERS, MCP.

TypeScriptMIT36
PageGitHub

Modality-agnostic IR for gate, neutral-atom, photonic-CV and pulse — FT-mode is one flag.

PythonApache-2.06
PageGitHub

Quantum-memory networks · 3 reproducible modes (Theoretical / Simulation / Hardware).

PythonApache-2.03
PageGitHub

macOS menu-bar workspace — 8 tools, 11 coding CLIs auto-detected, 12 automations.

SwiftMIT0
PageGitHub

Context engine for AI tooling — promote notes into Decisions, Sources, Tasks. MCP server.

TypeScriptMIT0
PageGitHub
Click a column header to sort. Search by name, language, licence, or any keyword in the tagline.
003Detailed view
05 records
01StableProductivity

Self-hosted project management with AI-agent governance.

A MIT-licensed issue tracker and work-graph control plane that feels like Linear, scales toward Jira parity, and keeps AI opt-in. Next.js 15 + PostgreSQL 16/pgvector + Redis 7 + LiveKit, Docker-first, localized in 30 languages.

  • v0.7.11 release line with IBM-modern app and email surfaces
  • 30-language localization with RTL Arabic and Hebrew
  • AI off by default: OpenAI, Anthropic, and native planner fallback
  • AGENTOWNERS, MCP write governance, approval gates, local runner beta
  • Docker Compose with Postgres 16, pgvector, Redis 7, optional LiveKit
  • Docker Hub update awareness and signed self-update handoff
Language
TypeScript
Licence
MIT
Status
Stable
Read moreView on GitHub →
02StableDeveloper tools

Your macOS menu bar, supercharged.

A small, fast, opinionated workspace that lives in the menu bar — tasks, focus, clipboard, notes, automations, and multi-provider AI, one click away. Local-first, no cloud account, plain JSON storage.

  • 8 tools in a single popover (⌘1–⌘8)
  • ⌘K command palette across tabs
  • Auto-detects 11 coding CLIs (Claude Code, Codex, Aider, opencode, Gemini, Amp, Goose, Continue, Plandex, Qwen, Ollama)
  • 12 one-click automations with approval gates
  • EN / TR localisation, 71 passing tests
  • macOS 14+, Swift 5.9, native SMAppService launch-at-login
Language
Swift
Licence
MIT
Status
Stable
Read moreView on GitHub →
03Coming soonDeveloper tools

Where context is built.

Turn your thinking into structured, queryable context that any AI tool can consume. Promote a sentence into a Decision, a paragraph into a Source, a checklist into Tasks — every object becomes addressable from Claude, Cursor, Copilot, or any MCP-compatible agent.

  • React 19 + TypeScript 6 + Tauri 2 — ~2 MB native bundle
  • Yjs CRDT for offline-first collaboration
  • Tiptap rich text + CodeMirror code blocks + LaTeX
  • Force-directed 3D knowledge graph (Three.js)
  • One-click export: CLAUDE.md, .cursorrules, AGENTS.md, schema/context.json
  • MCP server surface — agents query your workspace via standard protocol
  • 830+ tests, sub-100 ms cold start
Language
TypeScript
Licence
MIT
Status
Coming soon
Read moreView on GitHub →
04Public previewQuantum research

The quantum substrate of the Neura Parse stack.

One IR across gate, neutral-atom, photonic-CV and pulse modalities. Fault-tolerant promotion is one config flag (surface · BB qLDPC · Gidney–Shutty cultivation). Every run is an ed25519-signed manifest with offline-verifiable hash chains.

  • Modality-agnostic IR with deterministic content-hash
  • 157 passing tests across phases 1–6
  • PyMatching, sliding-window, BP+OSD, neural decoders
  • Pasqal Pulser, Strawberry Fields, MrMustard, Bloqade frontends
  • QutipEmulator, Aer, Stim, OpenPulse backends
  • SLURM / PBS connectors, crash-safe resume via signed lineage
  • Compose with Qiskit, Cirq, PennyLane — does not replace them
Language
Python
Licence
Apache-2.0
Status
Public preview
Read moreView on GitHub →
05BetaQuantum research

Quantum Memory-Augmented Neural Networks.

Hybrid quantum-classical neural networks with quantum memory operations. Reference implementation for quantum-inspired algorithms suitable for near-term NISQ devices, with three reproducible modes — Theoretical, Simulation, and Hardware.

  • Three modes: Theoretical (free), Simulation (free), Hardware (paid)
  • Q-Matrix quantum memory layer + Quantum-LSTM controller
  • Up to 20 qubits in simulation, 4–12 on real hardware
  • Qiskit 2.1+, IBM Quantum Network compatible
  • Cost estimator before any paid hardware run
  • Apache-2.0, codecov + CI, full docs
Language
Python
Licence
Apache-2.0
Status
Beta
Read moreView on GitHub →
004Contributing

Every project lives at github.com/neuraparse. Issues are reviewed weekly by the team that ships the project.

  • Bug reports — open an issue on the project repo with a reproduction.
  • Feature proposals — start a Discussion before sending a PR; we want the design conversation in public.
  • Security disclosures — email security@neuraparse.com, do not file a public issue.
  • Research collaboration — qmesh and QMANN are companion repos to peer-reviewed work; reach out for the Collaborator Edition or co-authored studies.
~/neura-parseQuickstart
# Try TaskNebula in 30 seconds
$ curl -fsSL https://raw.githubusercontent.com/\
    neuraparse/tasknebula/main/scripts/quickstart.sh | bash

# Install NeuraBar (macOS 14+)
$ git clone https://github.com/neuraparse/NeuraBar
$ cd NeuraBar && ./build.sh install

# Run a qmesh example
$ pip install -e ".[all]"
$ python3 examples/dag_cross_modality.py

# Hack on QMANN in simulation mode (free)
$ pip install -e .
$ python examples/02_simulation_mode.py
Build with us

Or hire us to integrate any of these into your stack — that's how the company keeps building.

GitHub →Talk to engineering