aegis (aegis-ai-suite) — Omniscience & Real-time Intelligent Operation Node
Orchestrate Projects with Intelligent Agents — Multi-Agent · Domain-Aware · Human-Governed · Enterprise-Integrated · Outcome-Driven
aegis connects your issue tracker (Jira, Backlog, Redmine, GitHub, GitLab), your Git repositories (GitHub, GitLab, Bitbucket Cloud & Data Center), and your AI models (OpenAI, Anthropic Claude, Google Gemini, Groq, GitHub Copilot CLI, Claude Code, Codex, Aider, Jira Rovo, or 100% local Ollama / LM Studio) in one unified workspace.
Click Start Fix on a ticket and a deterministic 8-stage pipeline of 13 Operational Agent Squads (120+ specialized agents) prepares the branch, synthesizes failing TDD tests, implements the code change in an isolated Git worktree, runs real runtime/lint/security gates (Semgrep, Trivy, Horusec), and presents a unified diff for human approval before anything reaches your branch.
⚡ 30-Second Quick Install & Launch (macOS · Windows CMD · Linux)
Install the single cross-platform Python package and launch the entire suite from your terminal:
# 1. Install the universal Python package from PyPI
pip install aegis-ai-suite
# 2. Launch Core CRM + Backend + Local Companion Agent + Standalone Chat
aegis-ai
# (or `aegis-crm`)
# 3. Or launch the dedicated Standalone Companion Chat directly
aegis-chat
Included Applications & Default Ports
| Application | CLI Command | Default URL | What It Runs |
|---|---|---|---|
| aegis Core CRM & Full Enterprise Suite | aegis-ai / aegis-crm |
http://localhost:8020 |
Full 13-Squad Multi-Agent CRM, Upstream AI Studio, Project Control Tower, Browser IDE, QA & Testing Center, 33 Built-in Reports, and Admin Governance |
| Standalone Companion Chat | aegis-chat |
http://localhost:5185 |
Distraction-free chat-first engineering desktop UI powered by Aegis Pinky (Rapid) & Aegis Brain (Deep Reasoning) with live tool streams and inline Git worktree diffs |
| Local Companion Agent | (Auto-started) | http://localhost:47771 |
Host-level Git worktree sandbox manager, AST code-graph engine, Playwright visual QA runner, and coding CLI bridge |
CLI Options (aegis-ai / aegis-workspace)
aegis-ai # Start Core CRM (:8020), Companion Chat (:5185) & Agent (:47771)
aegis-chat # Start & open Standalone Companion Chat (:5185)
aegis-ai --both # Open both Core CRM and Standalone Companion Chat in browser tabs
aegis-ai --no-browser # Run all servers headlessly without opening a browser window
aegis-ai --port 8020 --chat-port 5185 # Customize Core CRM, Chat, or Agent (--agent-port 47771) ports
aegis-ai --update aegis-ai-suite # Self-update to the latest release via pip
📊 Platform Scale at a Glance
| Metric | Scale | Scope |
|---|---|---|
| Native Integrations | 15+ | Jira, Backlog, Redmine, Confluence, GitHub, GitLab, Bitbucket, MS Teams, Slack, Outlook/M365, Entra ID, MCP & Cloud/Local LLMs |
| Operational Squads | 13 | L1–L6 corporate engineering hierarchy with dedicated Squad Managers & Directors |
| Autonomous Agents | 120+ | Specialized domain personas equipped with 107 real repository & workspace tools |
| Software Factory Stages | 8 | Deterministic quality gates from Spec Analysis & Red-State TDD to Security Audit & Merge |
| Built-in Reports | 33+ | Google TimesFM Predictive Forecasting, Cross-Tracker Analytics, AI Summarization & Dynamic Builder |
| Real Security Scanners | 3+ | Semgrep SAST, Trivy CVE/SCA, Horusec Secret Detection, Gitleaks & Bandit on every change |
| Interface Languages | 31 | Full i18n localization across English, Japanese (日本語), Chinese, Korean, Indic, European & RTL languages |
✨ Core Capabilities & Product Studios
1. 🏛️ Upstream AI Enablement & Solution Studio (For BAs, PMs & Architects)
Six dedicated upstream workspaces turn raw stakeholder workshops and business briefs into production-ready specifications:
- Meeting Minutes & Consensus Synthesizer: Converts raw transcripts into action items, architectural decisions, and Jira / Backlog / Redmine tickets.
- Requirements & BRD/BDD Structurer: Generates Functional Requirements (
機能要件), Non-Functional Requirements (非機能要件), Gherkin scenarios, and ambiguity alerts. - WBS & Man-Month Estimator: Computes Function Points, Story Points, and P10/P50/P90 Person-Days (
人日) & Man-Months (人月) with role-based cost projections. - Design Document Guardian: Audits Basic Design (
基本設計) and Detailed Design (詳細設計) for requirement traceability, concurrency, and SLA gaps. - Client Playback & Wireframe Prototype Generator: Builds interactive screen flows and records stakeholder sign-offs before coding starts.
- UAT Acceptance Evidence Pack & Partner Survey Builder: Exports traceability matrices and RFP/vendor evaluation surveys to CSV/Excel.
2. 🧑💻 Polyglot Coding Assistant & DevMate (For Developers)
- One-Click
Start FixDelivery Workflow: Branch → Plan → Red-State TDD → AST Implementation → Runtime Tests → Security Gate → Adversarial Diff Review → Human Approval → Commit & Tracker Sync. - DevMate Developer Companion: Pre-plans assigned tickets, heals failing tests and CI runs inside an isolated Git worktree, reviews local changes, and dry-runs pull requests.
- Fix from a Screenshot (Visual Bug-to-Code): Upload images, GIFs, or screen recordings — aegis extracts keyframes, runs English/Japanese OCR, maps the UI back to the exact source files, and proposes a verified fix.
- Bring Your Own Coding CLI or Local LLM: Run tasks through GitHub Copilot CLI, Claude Code, OpenAI Codex, Google Antigravity, Aider, or Jira Rovo, or run 100% offline/air-gapped with Ollama, LM Studio, vLLM, or llama.cpp.
- Living Code Graph & AI Code Guardian: Answer "What breaks if I change this symbol?" across callers, API routes, SQL tables, and tests, plus generate 10 comprehensive repository health & security reports.
- Browser IDE + Native IDE Extensions: Built-in Monaco editor, Git staging tree, and PTY terminal in the browser, alongside VS Code / Cursor, JetBrains, and Eclipse extensions.
3. 🧪 QA & Testing Center (For QA Engineers & SDETs)
- AI Test Case Generator Studio (BRD / HLD / LLD / Design Docs): Upload a BRD, High-Level Design (HLD), Low-Level Design (LLD), UI/UX Design Document (PDF, DOCX, PPTX, MD, OpenAPI), or verbal write-up. Select your AI Model, QA Specialist Agent, and Domain Skills to synthesize 8-vector test suites (Positive, Negative, Boundary, E2E Browser, API, Security, WCAG Accessibility, Performance).
- Right-Side Live Test Specification Preview: Inspect the generated Requirement Traceability Matrix, edit test steps inline, and export Markdown specifications or import directly into project test suites.
- 8 One-Click Ticket QA Actions: Analyze Issue, Generate Test Cases, Generate Test Data, Assess Blast Radius, Audit AC & BDD, Check History & Logs, Scaffold Tests, and Verify Coverage.
- Runnable Test Scaffolding & Visual Self-Healing Recorder: Generate executable Playwright (Python), PyTest, or Postman v2.1 suites, record browser flows across ~30 step types with Set-of-Mark visual AI selector self-healing, and generate PII-safe synthetic test data (including Japanese full-width / kana edge cases).
4. 🤝 Always-On Companions Mode, Standalone Chat (:5185) & Voice Director
- Standalone Companion Chat (
aegis-chaton:5185): Dedicated chat-first engineering app featuring the 2×2 Aegis Execution Matrix — Aegis Pinky (Rapid Execution) and Aegis Brain (Deep Multi-File Reasoning), live tool-call telemetry, and dirty-worktree resolution (Commit, Stash, or Clean). - Strategy Co-Pilot Manager Sentinel: Watches PR review bottlenecks, multi-sprint blockers, and SLA breach horizons; synthesizes daily async standups and weekly executive digests to Teams, Slack, or Outlook.
- Aegis Voice Director (
/aegis): Local neural speech-to-text (faster-whisper) and text-to-speech (Gradium Phonon 31M) for hands-free spoken workspace briefings and voice command orchestration.
🏢 The 13 Operational AI Agent Squads (120+ Specialized Personas)
| # | Operational Squad | Squad Leader Persona | Hierarchy | Primary SDLC Stage & Deliverables |
|---|---|---|---|---|
| 1 | Testing & QA Squad (10 agents) | Test Operations Director Agent | L5 Director | Stages 3, 6 & 7: BRD/HLD/LLD 8-vector test synthesis, Red-State TDD, Playwright E2E & flaky test RCA |
| 2 | Polyglot Coding Squad (12 agents) | Chief Software Architect Agent | L4 Principal | Stages 3–4: Syntax-gated AST patches across Python, Java, C#, TypeScript/React, Go & SQL in isolated worktrees |
| 3 | Code Review & Governance Squad (9 agents) | Code Review Squad Manager Agent | L6 Director | Stages 5–6: Semgrep SAST, Trivy CVE, Horusec secrets, cyclomatic complexity audit & adversarial PR review |
| 4 | Ticket Management & Intake Squad (8 agents) | Ticket Squad Manager Agent | L6 Director | Stage 1: Jira/Backlog/Redmine/GitHub intake triage, attachment OCR, INVEST validation & subtask breakdown |
| 5 | L1/L2 Support & Incident Squad (9 agents) | Director Support & Reliability Agent | L5 Director | 24/7 Triage: Stack-trace-to-commit RCA isolation, supervised rollback runbooks & post-mortem synthesis |
| 6 | Agile Ceremonies & Status Squad (8 agents) | Scrum Master Agent | L3 Lead | Continuous DORA: Async daily standups, Google TimesFM velocity forecasting & reviewer load balancing |
| 7 | ETL & Data Warehouse Squad (9 agents) | Chief Data Architect Agent | L4 Principal | Data Engineering: Kimball/Inmon star schemas, dbt/SQLMesh DAGs, column-level lineage & schema drift guards |
| 8 | Snowflake & DB Modernization Squad (8 agents) | Lead Snowflake DBA Agent | L3 Lead | DB Modernization: Cross-dialect SQL transpilation (Oracle/SQL Server/Teradata → Snowflake) & FinOps tuning |
| 9 | Cloud Delivery & DevOps Squad (10 agents) | Cloud Delivery Squad Manager Agent | L6 Director | Stage 8: Multi-cloud IaC (AWS/GCP/Azure), Docker/K8s manifests, blue-green/canary rollouts & squash merges |
| 10 | Travel & Airline Domain BA Squad (9 agents) | Principal Travel Domain Expert Agent | L3 Senior BA | Stages 1–2: Aviation FAR 117 crew legality, PNR lifecycle, IATA NDC / ONE Order, MRO & loyalty rule audits |
| 11 | Workspace Delivery Automation Squad (9 agents) | Workspace Governance Manager Agent | L6 Architect | Git Hygiene: Ephemeral worktree orchestration, dirty-tree resolution, monorepo guards & contract drift checks |
| 12 | SRE & Production Ops Squad (10 agents) | Incident Operations Squad Manager | L6 Director | Closed-Loop SRE: CloudWatch/Prometheus anomaly correlation & automated hotfix dispatch to Squads 2 & 3 |
| 13 | Product Management & Solution Squad (10 agents) | Product Management Squad Manager | L6 Director | Pre-Dev Studio: Idea-to-Epic/Story/Subtask synthesis, Hexagonal blueprints, DDL design & WBS cost estimates |
📈 33 Built-In Reports, TimesFM Forecasting & Enterprise Integrations
📸 Product Screen Gallery
 Project Details & Control Tower Unified Jira/Backlog/Redmine tickets with one-click Start Fix and 8-action AI Assistant |
 AI Ticket Assistant Analyze Issue, Generate Tests, Blast Radius, BDD Audit & Scaffold Tests on any issue |
 Visual Multi-Agent Workflow Studio 8-stage deterministic delivery pipelines edited on an interactive React Flow canvas |
 QA & Testing Center BRD/HLD/LLD AI Test Generator, live right-side spec preview, Playwright runner & traceability |
 Reports & TimesFM Predictive Analytics 33 built-in reports across TimesFM velocity forecasting, cross-tracker SLA & AI summaries |
 Standalone Companion Chat ( aegis-chat)Chat-first coding with Aegis Pinky & Aegis Brain, live tool streams & inline Git worktree diffs |
 Companion Inbox (DevMate & Strategy Co-Pilot) Always-on background branch preparation, CI healing, standups & human approval queue |
 AI Model Governance & Local LLM Registry Cloud LLMs, Ollama/LM Studio local models, Copilot/Rovo CLIs & token budget guardrails |
🏗️ System Architecture & 8-Stage Deterministic Pipeline
Real Pipeline Quality Gates (Zero Simulation)
Every agent run executes inside an isolated Git worktree (~/.nilswa-workspace-agent/worktrees/) and must clear four deterministic gates before human merge approval:
| Gate | Tools Executed by Local Companion Agent | Blocking Policy |
|---|---|---|
| 1. Runtime Gate | Auto-detected project suites: pytest, npm test / jest / vitest, go test, dotnet test, cargo test, mvn test, gradle test |
Blocks on any failing test suite ("No tests found" is marked skipped, never falsely passed) |
| 2. Lint & Type Gate | ruff, mypy, eslint, prettier, sqlfluff & AST syntax parser |
Blocks on syntax/parse errors and mypy type errors |
| 3. Security Gate | gitleaks, semgrep, trivy, bandit, pip-audit, npm audit + built-in secret/sink scanner |
Blocks on any finding at or above AEGIS_SECURITY_FAIL_ON (default high) in modified files |
| 4. Integration Gate | Real Git commit inside the isolated worktree + Human-in-the-Loop approval packet | Refuses commit if Runtime or Security gate failed |
🛠️ Technology Stack
| Layer | Technologies |
|---|---|
| Frontend Applications | React 18, Vite, Tailwind CSS, Monaco Editor, xterm.js, Recharts, React Flow, 31-Language i18n |
| Backend & Orchestration | Python 3.11–3.13, FastAPI, SQLAlchemy 2.0, LangChain, LangGraph, PostgreSQL Checkpointer, MCP Server/Client |
| Data & Vector Engine | PostgreSQL 16+ with pgvector & pg_trgm (Row-Level Security) + automatic local JSON store fallback |
| Local Companion Agent | FastAPI host daemon (:47771), Git worktree isolator, Tree-sitter polyglot AST engine (12 languages), Playwright Chromium |
| AI & Local Intelligence | Sentence-Transformers + Cross-Encoder Reranker (RAG), faster-whisper (STT), Gradium Phonon 31M (TTS), Google TimesFM, Tesseract OCR (EN/JA) |
📄 Author & License
- Creator & Principal Architect: Sachin Bose Pulikkathara (sachin.bose.p@gmail.com)
- GitHub: https://github.com/sachin-bose-p
- LinkedIn: https://www.linkedin.com/in/sachin-bose-p/
- License: Open Commercial & Internal Use (see root
LICENSE).
1 maintainer
Metadata
1 maintainer
Release files for aegis-ai-suite 1.0.1
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Actions Reset |
|---|---|---|---|---|
| aegis_ai_suite-1.0.1-py3-none-any.whl | Python 3 | none | any | Details for aegis_ai_suite-1.0.1-py3-none-any.whl |
Release files / aegis_ai_suite-1.0.1-py3-none-any.whl
| Download URL | aegis_ai_suite-1.0.1-py3-none-any.whl |
|---|---|
| Size | 81.1 MB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
3a1c9f3266a2db825523de6304468f87e46142e91d705474e529be1220521aad
|
|
BLAKE2b-256 checksum How to use checksums |
d088f137c4aa06eae199bf4eb816bf7db144535f9e787b2e3615fe329f35cbbf
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/7.0.0 CPython/3.12.15
|
Security for aegis-ai-suite 1.0.1
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
No provenance found
None of the files in this release were uploaded with provenance metadata.
A note on package security
PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.
To learn more about protecting yourself from malicious actors, visit our package security documentation.
