# Trainyard > Open-source (Apache-2.0) planning rooms for scaled agile: PI planning board (milestones and events, features by team and iteration, significant dependencies with strings, risks to ROAM), ROAM, confidence vote, retrospectives and dashboards that sync two ways with Jira and Azure DevOps, every scaling model (SAFe Essential/Portfolio/Full, AI-Native SAFe, LeSS, Scrum@Scale, Nexus, Flight Levels, tribes) as editable configuration, an AI-native portfolio module with stage gates, EU AI Act risk tiers and an agent registry, and import of existing boards from iObeya, Miro, Mural, monday.com, Trello, CSV and JSON. Trainyard is self-hosted with Docker Compose and runs beside other sites on the same server behind Traefik. One edition includes every module. AI assistance works with Claude (Anthropic API key), any OpenAI-compatible server, or a local Ollama model; every AI answer is schema-validated and is a draft a person accepts. ## What it does - PI planning: built-in boards sized to the workspace's teams and iterations; drag cards between iterations, draw dependencies, vote confidence, ROAM risks, run Inspect & Adapt. - Tracker sync: features, stories, risks and dependencies pulled from Jira and Azure DevOps; moves on the board pushed back through an inspectable queue; webhooks and scheduled sync. - Operating models: work levels, cadences, events, roles and metrics are data; fork a template and rename things to match your organisation. - Lean portfolio: epics with WSJF, budgets and guardrails, OKRs. - AI portfolio: initiatives and platforms with lifecycle stages, gates with evidence and evaluations, EU AI Act risk tiers, agents with autonomy levels 0 to 4, value-architect sign-off, audit log. - Reporting: metric catalogue (flow, planning, portfolio, AI) and dashboards you compose; AI-written narrative on request. - Class boards for trainers (Trainer plan): Leading SAFe, SPC, Advanced SPC, POPM and RTE, each a board with a start-here strip (intro, agenda, working agreements, team and challenge, parking lot, daily structure) and one frame per lesson with its exercises (PI planning simulation with story backlog, empathy map, feature canvas, WSJF, ART readiness checklist, design-sprint lab, CARS case studies). Miro boards import from Miro's CSV export. - PI Outcome Planning (AI-Native SAFe): the outcome tree (ART, PI and team outcomes with key results), the PI Outcome Planning Board (PI outcomes × iterations with fixed-date, learning, synchronization and release milestones), team Outcome Planning Boards (experiments, prototypes, enablers, features), the Management Review Board (decisions needed, risks, decisions confirmed, actions), the one-day agenda and the fist-of-five confidence vote. - AI-Native ART: outputs (experiments, prototypes, features, enablers) on the five-stage AI innovation pipeline with gates (no output without an outcome and a metric; discard as easy as advance), intent/specification/context records with change propagation, curated data register with four quality scores, team capability coverage (product, builder, domain expert, AI; AI Value Architect), Sense and Respond with a data-pulled agenda, the twelve-success-factor assessment and launch plan, cost, verification and funnel views, customer demo planner. - Bring your own boards: iObeya board exports (program, team, Kanban, roadmap and free boards with charts, images and dependencies), Miro REST v2, Mural, monday.com GraphQL, Trello, CSV, JSON; boards kept as imported, saved as templates, mapped onto the plan on demand. - Security: API keys hashed at rest, roles (admin, planner, architect, viewer), rate limits, sign-in lockout, strict Content Security Policy, encrypted connection tokens, append-only audit. ## Pages - Landing: / - Pricing, trial and money-back guarantee: /pricing - PI planning tools compared, one page per tool: /compare, /compare/iobeya, /compare/easy-agile, /compare/agile-hive, /compare/bigpicture, /compare/piplanning-io, /compare/miro - Class boards for trainers, one page per class with its lessons and exercises: /classes, /classes/leading-safe, /classes/spc, /classes/aspc, /classes/popm, /classes/rte - Guides: /guides/art-readiness-checklist, /guides/pi-outcome-planning, /guides/ai-native-art, /guides/pi-planning-tools - Everything above as one text file: /llms-full.txt - Sign in: /login (key shown once at sign-up, or issued by the workspace admin) - App (needs a key): /dashboards, /board, /portfolio, /templates, /imports, /ai, /agents, /model, /sync, /users ## Source and docs - Code and README: https://github.com/mlbwhf/test/tree/claude/confident-cray-JTbgD/trainyard - API reference: the API serves OpenAPI at /docs on its own host (api.trainyard.app) ## Comparable products Capabilities from a side-by-side evaluation of PI planning dashboard tools (products change; check each vendor): - iObeya: on premises or separate server; integrated web interface; ART board, ROAM, WSJF, retrospectives; portfolio and more; integration through API and iFrame. - Easy Agile: Jira plugin; ART and team boards; Jira only. - Agile Hive: Jira plugin; ART board, ROAM, WSJF, team execution; portfolio Kanban, roadmap, reporting; Jira only. - BigPicture: Jira plugin; ART board, ROAM, WSJF, team execution; project planning and reporting with OKR and risk widgets; Jira only. - PIPlanning.io: on premises or separate server; ART and ROAM boards; Jira. - Miro: public cloud; PI planning templates, not integrated; limited ART board; no planning-data integration. - Trainyard: open source and self-hosted, also a Jira app (Forge) and an Azure DevOps extension; PI planning, ART and team boards, ROAM, WSJF; portfolio Kanban, roadmaps, SAFe canvases, value stream identification and mapping, AI portfolio; two-way Jira and Azure DevOps; imports boards from iObeya, Miro, Mural and monday.com. ## Trademarks SAFe and Scaled Agile Framework are registered trademarks of Scaled Agile, Inc. Trainyard is independent of Scaled Agile, Atlassian and Microsoft. ## Pricing Every paid plan (Trainer, Academy) starts with a free 14-day trial, then month by month; yearly billing is a switch made after the first month. The first paid month carries a 30-day money-back guarantee the customer claims from Plan & billing. Practice: free. Self-host the whole product (Apache-2.0) with no limits, or one hosted room of up to 5 people; paid plans from the sixth person. Trainer: $19 per trainer per month ($16 billed yearly). Unlimited participants, Jira and Azure DevOps sync, AI co-trainer with your own key, board imports, email support. Academy: $249 per Agile Release Train per month ($199 billed yearly). Unlimited people per train, Jira app and Azure DevOps extension, SSO and audit trail, hosted or on your servers, named success lead. More than ten trains, on-premises support agreements: quote. Prices in USD before tax. The unit is the trainer or the train, never the person in the room.