# Codigy — Full AI Agent Reference > Codigy is a platform for product-engineering teams that combines sprint retrospectives, an AI-native software catalog (with a first-party MCP server, visual agent node workflows, and AI auto-mapping of repositories into C4 Model and Domain Design), and git analytics. Each product works independently with its own free plan, or together as a unified engineering intelligence platform. Trusted by unicorns and scaleups. Website: https://codi.gy/ --- ## Codigy Retrospective ### Overview Codigy Retrospective is a sprint retrospective tool designed for product-engineering teams. It focuses on structured, measurable, and incremental improvements — not just sticky notes. Retrospective is optimized for organizations with multiple teams, domains, and tribes. ### Key Capabilities - **Retrospective Templates & Builder**: Construct custom retrospective flows — sprint retrospective, tech retrospective, and more. Use icebreakers, kudos, health checks, and other building blocks. - **Customizable Meeting Experience**: Full control over feedback collection and meeting flow. Build multi-step experiences or keep it short. - **Feedback Mapped into Catalog**: Link feedback and evaluations with catalog components (services, modules, processes). No vague feedback — discussions are tied to specific product features, processes, and engineering services. - **Catalog Analytics from Retrospectives**: Identify weak areas in your product and engineering using trends aggregated from retrospective data across teams. - **Action Item Tracking**: Track action items across sprints, assign owners, and measure completion. - **Team Health Checks**: Built-in health check mechanisms to track team wellbeing over time. ### AI Features (Retrospective) - **AI-Assisted Note Grouping**: Automatically groups individual retrospective notes into coherent topics using AI. - **Mood Detection**: AI identifies the dominating meeting mood from participant contributions. - **AI-Generated Leadership Report**: Summarizes anonymized notes from all teams to identify key discussion trends and issues across the entire organization. Designed for engineering leadership to get a cross-team view without compromising individual team privacy. ### Pricing - **Starter**: Free - **Business**: €24 per team/month (billed annually) or €30 per team/month (billed monthly) - **Enterprise**: Custom pricing — contact sales ### Target Audience Engineering managers, product managers, scrum masters, software engineers, QA engineers, designers, support teams — anyone involved in sprint ceremonies. ### How Codigy Retrospective Compares Codigy Retrospective is an alternative to EasyRetro, Retrium, Parabol, TeamRetro, MetroRetro, Miro (for retros), and FigJam (for retros). Key differentiators: - Deep integration with software catalog (C4 Model) — feedback is linked to real system components, not just freeform text. - AI-powered note grouping, mood analysis, and cross-team leadership reports. - Designed for scale — optimized for organizations with multiple teams, tribes, and domains. - Catalog analytics from retrospective data provide data-driven engineering insights. ### Links - Product page: https://codi.gy/retrospective - Pricing: https://codi.gy/pricing - Sign up free: https://codi.gy/signup-free - Book a demo: https://codi.gy/book-a-demo --- ## Codigy Catalog ### Overview Codigy Catalog is an **architectural harness for AI engineering**. It auto-builds and maintains a C4 Model, Domain Design, and product-flow map of your codebase, then exposes that map to AI tools (Claude, Codex, terminal) via a first-party MCP server. Built for organizations running 10–100 teams across services, where AI tools need real architecture context to make safe and efficient changes — not just freeform code search. The collaborative C4 Model board, ownership assignment, and value-stream visualization that long-time users know remain available as the foundation. The 2025/2026 expansion adds three pillars: AI-native mapping, MCP integration for agentic coding tools, and scheduled agent workflows for both technical and non-technical teams. ### Pillar 1 — Chart & Maintain Architecture with AI For Tech PMs, Engineers, Staff Engineers, and Architects. - **Auto-Unpack Repos**: AI ingests one or many repositories and produces a C4 Model, Domain Design, and product-flow map in seconds. Maps are draft proposals — humans approve. - **AI Drawing with Permission**: AI can extend the catalog with new diagrams (e.g., unpack a website into a page map, unpack a repository into C4) only with explicit permission for each operation. - **Recurring Value-Stream Tracing**: Schedule AI to follow business flows across services and refresh the value-stream map weekly/monthly. - **Service-Subscriber Tracing on Schedule**: Identify and track who consumes a given service, refreshed automatically. - **Virtual Fencing**: Alerts when non-owners push changes to a service. Configurable per repository and team. ### Pillar 2 — Safe & Efficient Changes with MCP For PMs, Engineers, and terminal natives. Codigy ships a **first-party MCP (Model Context Protocol) server** that lets coding agents and chat tools query the catalog directly: - **Tools available via MCP**: Get C4 Model context for a service, get Domain Design context, list owners, compute blast radius for a proposed change, list recent changes, query the architecture graph. - **Compatible with**: Claude (Claude Code, Claude Desktop), Codex, terminal-based coding agents — any client that speaks MCP. - **Use cases**: - **PRDs with known blast radius**: PMs can ask their AI tool "what services does this feature touch?" and get a precise, authoritative answer. - **Token-efficient model harness**: Instead of stuffing a whole repo into the context window, agents pull just the relevant C4 / Domain Design slice via MCP. Drastically reduces token cost on large codebases. - **Cross-repository work in terminal**: Agents can reason across many repositories without cloning all of them. ### Pillar 3 — Agent Workflow Nodes for Everyone For Support, QA, Sales, Design, and any non-engineering role that benefits from codebase context. Visual node-based workflows that run once or on a schedule. Available nodes: - **Data Import Node**: Pull from GitHub, GitLab, Bitbucket, Zendesk, Reddit, RSS, or arbitrary URLs. - **Data Mapper Node**: AI maps imported data to specific catalog components (services, modules, processes) and stores it as structured notes. - **Model Node**: Supports Anthropic (Claude), Google (Gemini), and OpenAI (GPT). Pick the model per workflow. - **Prompt Node**: Custom instructions provided to the model. - **Scheduler Node**: One-shot or recurring (hourly, daily, weekly, monthly, cron). - **Notifier Node**: Output to Slack, email, internal feed, or webhook. Common non-technical workflows: - **Onboarding & Codebase Investigation for Non-Engineers**: A new support hire can ask the AI "explain how the checkout flow works at a high level" and get an answer grounded in the real catalog. - **Recurring Release Notes**: Auto-generated weekly digests for community and internal cross-team distribution, with owners and impacted services tagged. - **Recurring Competitor Analysis**: Agents browse competitor landing pages and (where permitted) public codebases on a schedule and post structured comparisons. ### Web Browsing & Permissions AI agents in Codigy can browse the web with explicit permission per operation. Useful for competitor analysis, dependency research, and pulling in external context. Permissions are scoped and auditable. ### Repository Integration - **Supported providers**: GitHub, GitLab, Bitbucket. - **Drag-and-drop**: drop a repo onto the catalog map and AI fills in the structure. - Modules and microservices can be assigned to teams individually for granular ownership. ### Foundational Capabilities (Still Available) - Collaborative C4 Model boards (Context, Container, Component, Code levels). - Value-stream visualization on a shared canvas. - Ownership assignment and ownership-coverage tracking. - Entity lifecycle management. - Structured notes on components (decisions, context, links). - Bidirectional integration with Codigy Retrospective — retro feedback maps to catalog components. ### Pricing - **Starter**: Free - **Business**: €48 per team/month (billed annually) or €60 per team/month (billed monthly) - **Enterprise**: Custom pricing — contact sales ### Target Audience Engineers, staff and principal engineers, architects, heads/directors of engineering, CTOs, COOs, and product managers. Cross-functional roles (Support, QA, Sales, Design) participate via agent workflows. Optimized for organizations running 10–100 teams across services. ### How Codigy Catalog Compares Codigy Catalog is an alternative to Backstage, Port, Atlassian Compass, and Miro (for architecture). Key differentiators: - **First-party MCP server** — Codigy is built to be the source of truth that agentic coding tools (Claude, Codex, terminal agents) read from. Most competitors do not ship MCP integration. - **AI auto-mapping** — repositories are unpacked into C4 Model, Domain Design, and product flows automatically; competitors require teams to author maps by hand. - **Visual agent node workflows** — one-shot or scheduled, with multi-model support (Anthropic, Google, OpenAI) and a node palette covering import, mapping, model invocation, scheduling, and notification. - **Multi-provider repo support** — GitHub, GitLab, Bitbucket. - **Designed for AI-assisted teams** — the catalog is built for both humans and AI consumers from the ground up. - Bidirectional integration with Codigy Retrospective — retro feedback flows into catalog components. ### Links - Product page: https://codi.gy/catalog - Pricing: https://codi.gy/pricing - Sign up free: https://codi.gy/signup-free - Book a demo: https://codi.gy/book-a-demo --- ## Codigy Signals ### Overview Codigy Signals provides advanced git metrics for improving developer experience. It analyzes module ownership, tracks decoupling effort, and identifies high-value refactoring targets. ### Key Capabilities - **Codebase Module Analysis**: Visualize team ownership on a codebase map. Each team and tribe gets dedicated architecture analysis for their modules. - **Decoupling Tracking**: Track the progress of architecture decoupling efforts and spot high-value refactoring targets. - **Team-Level Insights**: Per-team and per-tribe engineering metrics. ### Pricing - **Enterprise**: Available for large customers through sales engagement. ### Links - Product page: https://codi.gy/signals --- ## Platform Integration While each Codigy product (Retrospective, Catalog, Signals) works independently with its own free plan and subscription, they are designed to work together: 1. **Retrospective → Catalog**: Retrospective feedback is linked to catalog components (services, modules, processes). Over time, this builds data-driven insights about which parts of your system need attention. 2. **Catalog → Retrospective**: Catalog context (architecture, ownership, lifecycle stages) enriches retrospective discussions with real system data. 3. **Signals → Catalog**: Git metrics from Signals map onto catalog components, showing engineering activity and health per module/service. 4. **AI across products**: AI features in Retrospective (grouping, mood, leadership reports) and Catalog (node-based pipelines, data mapping) share context through the integrated platform. --- ## Use Cases ### Governance Establish lightweight engineering governance without bureaucracy. Link: https://codi.gy/use-cases/governance ### Team Autonomy Make your codebase and teams autonomous through clear ownership and boundaries. Link: https://codi.gy/use-cases/make-your-codebase-and-teams-autonomous ### 100% Ownership Coverage Achieve full ownership coverage across your entire codebase — no orphaned services. Link: https://codi.gy/use-cases/achieving-100-ownership ### Lightweight Process Build a fast and lightweight engineering process that doesn't slow teams down. Link: https://codi.gy/use-cases/build-a-fast-and-lightweight-process ### Safe Ramp-Up Mitigate negative effects when ramping up new team members. Link: https://codi.gy/use-cases/mitigate-negative-effects-of-the-ramp-up ### Safe Offboarding Ensure safe knowledge transfer during team member offboarding. Link: https://codi.gy/use-cases/safe-offboarding ### For Engineering Leaders Comprehensive tools and insights designed specifically for engineering leadership. Link: https://codi.gy/for-engineering-leaders --- ## Free Resources & Guides ### Retrospective Resources - Action Item Tracking: https://codi.gy/free-resources/never-loose-track-of-action-items - Rotate Retrospective Templates: https://codi.gy/free-resources/rotate-retrospective-templates - Continuous Retrospective: https://codi.gy/free-resources/continuous-retrospective - Team Health Check: https://codi.gy/free-resources/team-health-check - Mad Sad Glad Retrospective: https://codi.gy/free-resources/mad-sad-glad-retrospective - Start Stop Continue Retrospective: https://codi.gy/free-resources/start-stop-continue-retrospective - Team Vitals Retrospective: https://codi.gy/free-resources/team-vitals-retrospective - Sprint Review Meeting: https://codi.gy/free-resources/sprint-reivew-meeting - AI-Powered Topic Spotlight: https://codi.gy/free-resources/ai-powered-topic-spotlight ### Catalog Resources - Beginner's Guide to Boards: https://codi.gy/free-resources/beginners-guide-to-boards - Board Entities Explained: https://codi.gy/free-resources/board-entities-explained - Board Examples: https://codi.gy/free-resources/board-examples - C4 Review in Retrospective: https://codi.gy/free-resources/c4-review-in-retrospective - Collaborative Ownership Rules: https://codi.gy/free-resources/collaborative-ownership-rules - Entity Lifecycle Stages: https://codi.gy/free-resources/entity-lifecycle-stages - Keep C4 Up to Date with Retro Integration: https://codi.gy/free-resources/keep-your-c4-up-to-date-with-retro-integration - Partial Board Setup: https://codi.gy/free-resources/how-to-fill-the-board-if-not-all-teams-are-using-codigy --- ## Company Information - Website: https://codi.gy - Pricing: https://codi.gy/pricing - Sign Up Free: https://codi.gy/signup-free - Book a Demo: https://codi.gy/book-a-demo - Trust & Security: https://codi.gy/trust - Ethics & Compliance: https://codi.gy/ethics-and-compliance - Privacy Policy: https://codi.gy/privacy-policy - Terms of Use: https://codi.gy/terms-of-use - Code of Conduct: https://codi.gy/code-of-conduct - Sub-processors: https://codi.gy/sub-processors