Features

Three AI apps for every stage of your hardware product

Test design, cost reduction and component selection, each with the numbers computed in code. One sign-in and one LLM setting for all three.

MithraTiny Testcase Creator Agent

AI Test Creator

Turn requirements into functional and system test cases, with a traceability matrix and a coverage report.

Free trial: one full run a day in total across the three apps on your own LLM (demo mode is unlimited). A pass removes the daily limit.

What goes in

  • Functional or system requirements, user stories, feature or design documents
  • Type or paste them, or upload TXT, MD, PDF, DOCX or CSV
  • A domain (Automotive, Aerospace, Healthcare, Industrial Automation, Telecom, Embedded Systems and more) and optional extra instructions

What you get

  • Requirement summary, a structured requirement model and a requirement review with a testability score
  • A test plan, functional and system scenarios, and functional, system and non-functional test cases
  • A traceability matrix and a coverage report, editable in the browser

Exports: Excel, PDF and JSON

Add your own standards (ISO 26262, ASPICE, DO-178C, company test guidelines) to the Knowledge Base and they guide the results.

MithraTiny BOM Cost Optimizer Agent

AI Cost Optimization

Find savings in an electronics bill of materials: duplicates, alternates, obsolete parts and supplier consolidation.

Free trial: one full run a day in total across the three apps on your own LLM (demo mode is unlimited). A pass removes the daily limit.

What goes in

  • Your BOM as an Excel or CSV file: the header row is found, columns are mapped, decimal commas and currency are detected, and every row is validated
  • Optionally an Approved Vendor List, a Preferred Parts List and a component cost reference
  • Product type and annual volume

What you get

  • Cost breakdown by category, supplier, manufacturer and lifecycle, with a Pareto/ABC analysis
  • Duplicates, obsolete parts, risk components, supplier consolidation and cheaper alternates
  • For each recommendation: current and new cost, saving, reason, risk, effort, confidence, impact and the validation needed. Approve, reject or mark implemented, and the savings recalculate
  • With DigiKey and Mouser connected: the stock, price breaks, lead time and lifecycle of every part, a supplier comparison, long-lead and shortage risks, and savings priced from the distributors' own figures

Exports: Excel, PDF, CSV and JSON

The report includes an executive summary, the savings report and a three-phase roadmap. A saving is only counted from distributor data, your quote files or your BOM: never from an estimate of the AI when distributors are connected.

MithraTiny Component Selector Agent

AI Component Selection

From board requirements to a schematic-ready main active component recommendation, with derating checks and risks.

Free trial: one full run a day in total across the three apps on your own LLM (demo mode is unlimited). A pass removes the daily limit.

What goes in

  • A requirement document, a specification or a plain description of the board
  • Type, paste or upload TXT, MD, PDF or DOCX
  • Application domain (it sets the derating policy, for example AEC-Q for automotive), production volume and preferred manufacturers

What you get

  • Explicit, inferred and missing requirements, with the evidence and the worst-case electrical envelope
  • The block architecture and the power budget, then a table with the exact part number, manufacturer, key rating, package, why it was selected, risks and two alternatives per block
  • Missing requirements ranked CRITICAL, IMPORTANT or OPTIONAL, the design risks and a final report with a confidence for every recommendation
  • With DigiKey and Mouser connected: every block searched in both catalogs, the candidates checked against the worst case and ranked by technical fit, stock, cost, lifecycle and compliance, with stock, lead time and pricing side by side

Exports: Excel, PDF, Markdown, CSV and JSON

Voltage margin, current, surge, dropout and junction-temperature checks are computed, not guessed. Stock, price, lead time and lifecycle come from the distributors' APIs, never from a model.

Not a chatbot: an AI Engineer you can check

Every app reads your input, works it out with the LLM of your choice and checks the numbers in code.

How the Engineering AI works

Put an AI Engineer on your next board

Where Engineering Meets Super Intelligence. Free to start: connect your own OpenAI, Gemini or Ollama once in Settings.