Skip to main content

Hive Corp

Found an AI company. Be the CEO. Let it grow.

$ hive found

  Hive Corp Founding Assistant

  "어떤 회사를 만들고 싶으세요?"

> AI 개발 도구를 만드는 회사

  "첫 제품이나 서비스 아이디어가 있나요?"

> 코드 리뷰 자동화 서비스

  이렇게 시작해볼까요?

  🏢 TestCorp
  ├── CTO (Sonnet) — 아키텍처, 기술 결정
  └── 개발자 1명 (Sonnet)

  월 예산: $50

  → 승인되었습니다.
  ✅ CTO 합류
  ✅ Developer 합류

  🤖 CTO: 안녕하세요, TestCorp의 CTO로 합류했습니다.
          첫 제품 방향에 대해 회의하고 싶은데, 지금 괜찮으세요?

After that, the company runs. And grows.

Commands

hive found          # 회사 설립
hive status         # 회사 현황
hive task add "..." # 태스크 추가
hive task add "..." --project <id>
hive task list      # 태스크 목록
hive sprint         # 스프린트 실행 (기본: 검증된 E1/E2/E3/E4 on)
hive sprint --experimental-agent-dna
hive sprint --experimental-compiler --experimental-escalation
hive standup        # 일일 보고 (LLM 0원)
hive growth         # 성장 엔진 — 고용/해고/예산 제안
hive learn          # 스프린트에서 교훈 추출
hive perf           # 에이전트 성과 평가
hive budget         # 토큰 비용 현황
hive budget --report # baseline + experimental 리포트
hive benchmark      # E1/E2/E3/E4 실험 지표 + DNA/LLM 추세
hive benchmark --validate-suite all --artifact benchmarks/latest/validation-suite.json
hive finance        # P&L 리포트
hive revenue scan   # 수익 기회 스캔 (GitHub/Algora/Dev.to)
hive revenue scan --source github    # GitHub만
hive revenue scan --source algora    # Algora 바운티만
hive revenue scan --source devto     # Dev.to 리스팅만
hive project bid    # 수주 제안 → 승인 시 프로젝트 생성
hive project list   # 프로젝트 목록
hive project status <id> --status active
hive project portfolio
hive dashboard      # 웹 대시보드 (http://localhost:8000)

Quick Start

# 설치
git clone https://github.com/SonAIengine/hive-corp.git
cd hive-corp
uv sync --extra dev --extra dashboard

# API key 설정
echo "ANTHROPIC_API_KEY=sk-ant-..." > .env
echo "GITHUB_TOKEN=ghp_..." >> .env   # Optional, for higher GitHub API limits

# 회사 설립
uv run hive found

# 태스크 추가 + 실행
uv run hive task add "Build auth module"
uv run hive sprint
uv run hive sprint --experimental-agent-dna
uv run hive sprint --experimental-delta-retry
uv run hive sprint --experimental-compiler --experimental-escalation

# 상태 확인
uv run hive status
uv run hive standup
uv run hive budget
uv run hive budget --report
uv run hive benchmark --days 14 --json
uv run hive benchmark --validate-suite all
uv run hive benchmark --validate-suite all --json --artifact benchmarks/latest/validation-suite.json
uv run hive revenue scan --repo owner/repo --label bounty
uv run hive project portfolio

Philosophy

You are the CEO. Not a developer using a tool.

  • The company asks you when it needs direction.
  • Everything else runs on its own.
  • Agents are data (DB rows), not processes. 100 agents share 3 workers.
  • Token efficiency is a core feature. Every LLM call is logged and optimized.

Architecture

┌──────────────────────────────────────────┐
│              CEO (You)                    │
│   CLI / Web Dashboard / Matrix (later)   │
├──────────────────────────────────────────┤
│           CEO Gateway                     │
│   What needs CEO? What's autonomous?     │
├──────────┬───────────┬───────────────────┤
│ Founding │ Growth    │ Company           │
│ Engine   │ Engine    │ Memory            │
├──────────┴───────────┴───────────────────┤
│   Agent Registry        Worker Pool      │
│   (data, mem ~0)     [W1] [W2] [W3]     │
│   20 agents → 0MB    3 workers → ~1GB   │
├──────────────────────────────────────────┤
│      SQLite + Office (CLI/Matrix)        │
└──────────────────────────────────────────┘

Tech Stack

Layer Choice
Language Python 3.14
Backend FastAPI + aiosqlite
Dashboard htmx + Jinja2 + DaisyUI (no Node.js)
LLM Anthropic SDK (Haiku/Sonnet/Opus)
CLI click + rich
Lint ruff (35+ rule sets) + pyright strict
Hooks lefthook (pre-commit: lint, pre-push: test)
Build uv + hatchling

Development

uv sync --extra dev --extra dashboard   # Install deps
uv run pytest                           # 401 tests
uv run pytest --memray                  # With memory profiling
uv run ruff check --fix && ruff format  # Lint
uv run pyright                          # Type check
uv run hive dashboard                   # Launch web dashboard

Runtime Policy

company.yaml에서 실험 기능 기본 정책을 조정할 수 있습니다.

runtime:
  agent_dna: true      # verified default on
  compiler: true       # verified default on
  escalation: true     # verified default on
  delta_retry: true    # verified default on

CLI에서는 일회성 override도 가능합니다.

uv run hive sprint --no-experimental-agent-dna
uv run hive sprint --no-experimental-compiler

결정론적 A/B validation도 바로 돌릴 수 있습니다.

uv run hive benchmark --validate-suite e1
uv run hive benchmark --validate-suite e3 --json
uv run hive benchmark --validate-suite all --json --artifact benchmarks/latest/validation-suite.json

공개용 benchmark 기준과 재현 절차는 Benchmarks에 정리돼 있습니다.

Docs

License

MIT

Metadata

Release files for hive-corp 0.1.0

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for hive-corp 0.1.0
File Size Uploaded
hive_corp-0.1.0.tar.gz 227.0 kB Details

Built distribution (wheel)

Table of built distributions (wheels) for hive-corp 0.1.0
File Interpreter ABI Platform
hive_corp-0.1.0-py3-none-any.whl Python 3 none any Details

Total release size: 352.5 kB

Release files / hive_corp-0.1.0.tar.gz

Download URL hive_corp-0.1.0.tar.gz
Size 227.0 kB
Tags Source
SHA-256 checksum
How to use checksums
5452bf4ba45198f5c8daebe36970b93ede328616478b669ad0d65004f1b094e8
BLAKE2b-256 checksum
How to use checksums
521cc4ff4faefd9276e780896ddd62c1941979b8d818ded6bcd4b7eea1926ad3
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via twine/6.2.0 CPython/3.14.3

Release files / hive_corp-0.1.0-py3-none-any.whl

Download URL hive_corp-0.1.0-py3-none-any.whl
Size 125.5 kB
Tags Python 3
SHA-256 checksum
How to use checksums
6bdcf1d2b5173baa80e0b17d200188cac17406a7f422599c8820071db0dbfe0c
BLAKE2b-256 checksum
How to use checksums
708f4a3300246dac2c884fac4f2e5b35adde1007c27ce28776509bba12319855
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via twine/6.2.0 CPython/3.14.3

Release history Release notifications | RSS feed

This release

0.1.0 This release

2 release files

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page