Spark / FAQ
Spark FAQ
Is it right for us
Can Spark turn our validated idea into a working MVP?
Yes. You get a working prototype that can serve as your MVP, with the requirements, design decisions and handoff needed to turn it into a product. Spark is Mantle's framework and certification for disciplined vibe coding: building software with AI assistance in a structured way.
Can we take our Litmus results straight into a Spark build?
Yes. Your Litmus work carries straight in. Spark brings in the hypothesis, customer evidence and kill criteria, and builds a working MVP, not a throwaway demo. If your idea still needs validating, start with Litmus. Spark also includes certification for engineers, so the skill stays in your team.
Will Spark build a skill across our team, not just one developer?
Yes. Spark builds a team capability, not single-developer heroics. The programme includes certification in the discipline, with a curriculum, a capstone and a recertification cycle, so the way of working stays with your team.
How it works
Will Spark use the right AI model for each part of the build?
Yes. Each task gets a model chosen for it, not a default. Builds use structured prompts, test scaffolds, automated evaluation and architecture decision records. Tests are built into the prototype, and guardrails watch for duplicated code, complexity and dependency problems.
Can we check the quality of what the AI builds as we go?
Yes. Quality is tested from the start. An evaluation harness uses an AI evaluator alongside fixed test assertions, and trace replay and prompt audit are built in, so you can see how the work was produced.
Will our product requirements be written down as Spark builds?
Yes. Your requirements stay on record. Forge, Mantle's product management and delivery tool, automates the product documents Spark runs on, and Spark keeps architecture decision records as it builds.
Your systems and data
Can the Spark prototype grow into our full product?
Yes. The prototype is built to grow into your product. Its data model, sign-in model, monitoring and deployment path are designed for that from the start. It then hands off to Lightning Build, Mantle's delivery program, with its tests and architecture record intact.
Will the prototype come with tests our team can keep using?
Yes. Your tests come with the prototype. Unit, integration and behavioral tests are scaffolded into it, and the test harness travels with it into Lightning Build.
Getting started
Can we talk to your team about a Spark build?
Yes. Use the form on our contact page, and our team will be in touch.
Do we bring a validated idea to start a Spark build?
Yes. Spark starts from evidence: a validated idea with its hypothesis, customer evidence and kill criteria. If your idea still needs that proof, Litmus, Mantle's idea validation workshop, is the place to start.
Pricing and support
Can we get a price for Spark that fits our scope?
Yes. Pricing depends on scope. Use the contact form and our team will be in touch.
Will the team that runs our MVP get what it needs to operate it?
Yes. The handoff to Lightning Build, Mantle's delivery program, includes a clean test harness, the architecture record and the operating posture, so the team taking it forward starts informed.