Waydev SDLC 15B is our own AI model, trained on software delivery data. Ask questions in plain English, generate dashboards from a sentence, and turn any widget into an explanation, a forecast, or a full dashboard.

Deployed on-premise, it runs on your own GPU, so your prompts, data, and answers never leave your network. *

Docs: https://docs.waydev.co/v5.0/docs/waydev-sdlc

Highlights

• Generate dashboards. Describe what you want to track, for example "Build a dashboard on AI tool adoption and merge survival", and Waydev SDLC 15B builds it for you, complete with the right metrics and widgets.
• Create dashboards from any widget. Click the sparkle icon on any widget and choose Create related dashboard to expand a single chart into a full dashboard on the same topic. Docs: https://docs.waydev.co/v5.0/docs/dashboards
• Explain, Predict, Improve. Every widget now has three AI actions. Explain tells you what the insight means. Predict forecasts where the metric is heading. Improve gives you actionable recommendations.
• Ask your data. Ask questions like "Why did cycle time go up last sprint?" and get answers grounded in your PRs, commits, issues, deployments, teams, and AI usage. No analyst or custom report needed. Docs: https://docs.waydev.co/v5.0/docs/waydev-agent
• Private by design.* Waydev SDLC 15B can run on a GPU server inside your network. It makes no calls to outside AI vendors and keeps working without internet access once installed.

Ask Waydev

• Generate widgets from a description, for example "Show me a widget with the most used AI tools".
• Download generated widgets, add them to an existing dashboard, or turn them into a related dashboard straight from the chat.
• Get explanations of metric changes and forecasts of trends.
• Get help choosing and understanding Waydev metrics, including DORA, PR cycle time, review depth, investment allocation, and AI coding adoption.
• Generate valid SQL on the Waydev schema.
• Get how-to answers from the Waydev documentation.
• Get PR reviews and help with coding questions.

Privacy and Compliance

• Local execution.* Prompts, data, and answers stay inside your infrastructure.
• Existing permissions. The model reads data only through Waydev's tools, which apply your existing account and permission checks.
• Works Council mode. Answers stay at team level, and the model declines to rank or single out individuals.
• Anonymization. Colleagues' names can be hidden in answers.
• Audit logging. Every AI request is logged with the model used, the skill applied, and whether a customer-owned key was used.

Built on Waydev Expertise

• Trained on more than ten years of Waydev's work: real questions from support, onboarding, and executive reviews, plus our metric definitions and data model.
• Every training answer is checked automatically. Numbers must match reference SQL, and answers with invalid SQL, invented columns, or leaked names are discarded.
• A new version ships only if it beats the current model on a frozen benchmark of more than 100 cases.
• 9 out of 10 answers are correct, with 53% fewer wrong answers than the open-source base model.

Deployment

• Runs on a single NVIDIA GPU: 16 GB minimum (64K context, 4 parallel requests) or 48 GB recommended (128K context, 6 parallel requests).\*
• Infrastructure cost is fixed, with no per-token fees.\*
• Waydev SDLC 15B is the default model. You can also use Claude or bring your own Anthropic or OpenAI key.
• Custom builds are available, covering your terminology, KPIs, internal docs, workflows, and works agreement rules.

* Applies only to on-premise deployments, where Waydev SDLC 15B runs on a GPU server inside your own network. These capabilities are not available on Waydev Cloud. Contact your Waydev account manager to discuss on-premise deployment.

References:

https://docs.waydev.co/v5.0/docs/waydev-agent
https://docs.waydev.co/v5.0/docs/waydev-sdlc
https://docs.waydev.co/v5.0/docs/ai-providers-byok
https://docs.waydev.co/v5.0/docs/dashboards