TL;DR
Traceformer.io offers an AI-driven reviewer for PCB schematics that analyzes design files and datasheets to surface errors before fabrication. The service uses a multi-agent pipeline, supports KiCad and Altium netlists, and is available in Free, Hobby and Pro plans with transparent token-based pricing.
What happened
Traceformer.io published an updated tool that runs automated, LLM-powered checks against PCB schematics using project files and component datasheets. The system ingests a complete KiCad project or an Altium netlist, breaks the design into subsystems, and issues a work plan for review. That plan is executed by up to ten parallel worker agents that search datasheets and assemble evidence, and a final merger agent synthesizes the findings into structured output showing errors, warnings, verified items and missing information. The site highlights configurable review rules, automatic retrieval of datasheets from links or trusted searches, and the ability to pick model providers such as OpenAI or Anthropic. Traceformer advertises transparent token usage and cost estimates, included API usage per month, and billing through Stripe, with tiers ranging from a free single-review plan to paid Hobby and Pro subscriptions.
Why it matters
- Aims to reduce fabrication respins by finding schematic mistakes earlier in the design cycle.
- Combines schematic parsing with datasheet evidence to flag issues beyond standard ERC checks.
- Parallel, specialist worker agents are intended to speed reviews while controlling model cost.
- Multi-provider and multi-model support gives teams flexibility in choosing underlying LLMs.
Key facts
- Uses a three-phase orchestrator pipeline: planner, parallel workers (up to 10), and a merger.
- Planner parses KiCad projects or Altium netlist files and identifies circuit subsystems (power, comms, sensors, etc.).
- Workers search up to 40 datasheets per project (limits vary by plan) and use lightweight models to reduce cost.
- Final merger agent synthesizes evidence into structured findings, including errors, warnings, verified items and missing info.
- Supports multiple model providers, specifically citing OpenAI and Anthropic, and lists models such as GPT-5.2, Opus 4.5 and Sonnet 4.5.
- Automatic datasheet retrieval uses provided links or trusted searches to gather specifications.
- Configurable review parameters include token limits, design rules and other settings.
- Pricing tiers: Free (one review/month, up to 10 datasheets), Hobby ($10/month, up to 20 datasheets), Pro ($20/month, up to 40 datasheets); each plan includes some API usage and a platform fee structure.
- Billing and payments are handled via Stripe; the site emphasizes transparent token usage and cost estimates.
What to watch next
- How the tool's findings compare in practice to traditional ERC/DRC workflows and whether teams adopt it broadly (not confirmed in the source).
- Updates to the supported model lineup and how new model integrations affect review quality and cost — the site says models are updated as new models become available.
- Privacy, data handling and hallucination-mitigation measures for customer projects and datasheets (not confirmed in the source).
Quick glossary
- LLM: Large Language Model — a machine learning model trained to understand and generate human-like text; used here to analyze design files and documentation.
- Schematic: A diagram that represents the electrical connections and components in a circuit.
- Datasheet: Manufacturer-provided documentation that specifies component electrical characteristics, limits and recommended usage.
- Netlist: A textual representation of a circuit that lists components and the nets (connections) between their pins; commonly exported from PCB design tools.
- ERC/DRC: Electrical Rule Check / Design Rule Check — automated checks performed by CAD tools to find basic schematic and layout issues.
Reader FAQ
What results can I expect from a Traceformer review?
not confirmed in the source
How does Traceformer use customer data?
not confirmed in the source
How does the service prevent hallucinations in its findings?
not confirmed in the source
Does this replace traditional ERC/DRC checks?
not confirmed in the source
Traceformer Benefits How it Works Features Pricing FAQ Open App Sign in Catch schematic mistakes before fab. AI-driven electrical design checks, powered by your datasheets and design files. Try it…
Sources
- Show HN: An LLM-Powered PCB Schematic Checker (Major Update)
- An LLM-Powered PCB Schematic Checker (Major Update)
- Testing Generative AI for Circuit Board Design
- A Large Language Model Powered Integrated Circuit …
Related posts
- DoorDash bans driver after alleged AI-generated fake delivery photo
- Ripple: a daily puzzle game that traces second- and third-order effects
- Agentic Patterns: Curated Catalogue of Repeatable, Traceable AI Agent Patterns