Skip to content

Releases: ARCHITECTURA-AI/AIdeator

v0.9.9 — Production Readiness, Type Safety & Engine Stabilization

30 Apr 08:05

Choose a tag to compare

🚀 AIdeator v0.9.9

AIdeator is a complete idea validation system that takes you from raw idea to build decision — with scored desk research, structured user interviews, and behavioral experiments. Stop overthinking. Start validating.

What's new in v0.9.9

  • Production-Ready Dependency Architecture: Standardized the dependency tree and added Mypy stubs for toml, PIL, and SQLAlchemy, achieving zero type errors across 61 source files.
  • 100% Code Quality Compliance: Achieved a completely green pipeline for both ruff and mypy with a strict 100-character line limit across the entire repository.
  • Stabilized Engine Core: Hardened the orchestrator's event state transitions, eliminating race conditions in event broadcasting and ensuring deterministic run status tracking.
  • Adversarial "Battle Mode" Verification: Refined the Bull/Bear reasoning engine with production-ready prompt templates for high-fidelity adversarial synthesis.
  • Registry Hardening: Rebuilt the Search and LLM registries with strict error handling, dynamic API key resolution, and fallback provider logic.

Install / Upgrade

pip install --upgrade aideator

Quick Start

# Initialize your local setup
aideator config init

# Run a local-only validation
aideator validate "A solar powered desk lamp" --mode local-only

# Start the command center
aideator serve # → http://localhost:8000

Privacy Modes

Mode External Calls
local-only None (Uses Ollama/Local LLMs)
hybrid Keyword-only (Encrypted search)
cloud-enabled Full context (OpenAI/Anthropic)

🔗 PyPI: https://pypi.org/project/aideator/

Verification Checklist (v0.9.9)

  • ruff check . -> PASS
  • mypy . -> PASS (61 files)
  • pytest -> GREEN
  • Engine State Machine -> STABLE

AIDEATOR_OS v0.9.9 is now online and ready for intelligence extraction.

v0.9.8 — Stability, Security & CLI Hardening `Latest`

29 Apr 05:21

Choose a tag to compare

🚀 AIdeator v0.9.8

AIdeator is a local-first AI idea validation engine that scores your startup ideas on demand, competition, and risk — entirely on your machine.

What's new in v0.9.8

  • CLI Robustness: Fixed Windows-specific module collisions and naming conflicts. The CLI now intelligently re-prioritizes local packages to ensure a smooth "out-of-the-box" experience on all OSs.
  • Authentication Hardening: Implemented byte-aware password truncation (72-byte limit) to satisfy bcrypt constraints, ensuring stability even with long multi-byte passwords or emojis.
  • CI/CD Stability: Fully isolated test database lifecycles, eliminating "database is locked" errors and race conditions in the GitHub Actions pipeline.
  • Package Flexibility: Added python -m aideator support via the new __main__.py entry point.
  • Dependency Cleanup: Pinned critical libraries (like bcrypt < 4.0.0) to ensure 100% compatibility with passlib and other core security components.
  • Demo Mode Fix: Resolved path-resolution issues for the /demo endpoint, ensuring the demo data loads correctly in all environments.

Install / Upgrade

pip install --upgrade aideator

Quick Start

# Initialize your local setup
aideator config init

# Run a local-only validation
aideator validate "A solar powered desk lamp" --mode local-only

# Start the command center
aideator serve # → http://localhost:8000

Privacy Modes

Mode External Calls
local-only None (Uses Ollama/Local LLMs)
hybrid Keyword-only (Encrypted search)
cloud-enabled Full context (OpenAI/Anthropic)

🔗 PyPI: https://pypi.org/project/aideator/


Verification Checklist (v0.9.8)

  • ruff check . -> PASS
  • mypy . -> PASS
  • pytest -> GREEN (Isolated DBs)
  • CLI native execution -> STABLE

AIDEATOR_OS v0.9.8 is now online and ready for intelligence extraction.

v0.9.7 — Stability & Provider Enhancements

27 Apr 10:52

Choose a tag to compare

🚀 AIdeator v0.9.7

AIdeator is a local-first AI idea validation engine that scores your startup
ideas on demand, competition, and risk — entirely on your machine.

What's new in v0.9.7

  • Improved provider abstraction for LLMs (Ollama, OpenAI, Anthropic, Mistral)
  • Enhanced HTML report generation via Jinja2 templates
  • Better scoring benchmarks against 12 reference SaaS products
  • Bug fixes and stability improvements

Install / Upgrade

pip install --upgrade aideator

Quick Start

aideator config init
aideator serve # → http://localhost:8000

Privacy Modes

Mode External Calls
local-only None
hybrid Keyword-only
cloud-enabled Full context

🔗 PyPI: https://pypi.org/project/aideator/