Violet Labs

About three years in at Violet Labs, I started to realize there was another term for what we were building. I found myself dancing around it in pitches and demos. Violet... "shepherds your product throughout its lifecycle" "manages your product's data from cradle to grave" "connects and orchestrates product data end-to-end". We were building Product Lifecycle Management. But, our platform is entirely different from what most people associate with PLM (ie. MCAD management. with a sprinkling of ECAD). We are solving for connectivity first - the piece that most every PLM ignores and tries to tack on at the end. So we're at once a PLM, and very much not at all a PLM. Maybe even... anti-PLM. It's a term that carries significant baggage - both positive and negative. It's been exciting to see feedback from industry titans like Doug Macdonald who are recognizing the value of Violet in this broader sense.

Agentic AI for hardware engineering has a plumbing problem. The agents can be smart. But if the data is fragmented across PLM, CAD, ERP, requirements, Slack, email, manufacturing systems, and supplier tools, they are still flying blind. Lucy Hoag explained at #ThreadedMiami that this is the problem Violet Labs is attacking. Their framing is direct: Hardware products are becoming more complex, more multidisciplinary, and more AI-enabled. But the way we build has not meaningfully changed since mechanical-centric PLM became dominant in the 1990s. The missing layer is not another app. It is knowledge orchestration. Violet positions itself as infrastructure for connected build: → 75+ no-code integrations → Hub-and-spoke data framework → Ontology-structured data → Universal access via MCP → Declarative sync handlers → Round-trip auditability → Monitoring and observability → Workflow triggers and approval gates → Context from relationships, metadata, decision history, Slack, and email That is the part of agentic AI most demos skip. Agents need permissions. They need source-system context. They need clean mappings. They need audit trails. They need to trigger workflows without destroying traceability. They need to know when two systems disagree. Violet’s line is basically this: Without connected data, agents are blind to system boundaries and data inconsistencies. With Violet, agents operate on a clean, connected, real-time data graph. That is not a flashy copilot message. It is the infrastructure layer hardware AI actually needs. More startup intelligence at https://threadmoat.com #BetterCallFino #AIAcrossTheProductLifecycle #Threaded #ThreadMoat #AgenticAI #DigitalThread #HardwareEngineering #PLM #MCP #EngineeringData #AIInfrastructure

I'm not a regular Chat, I'm a cool Chat ✨ We're excited to launch Violet Chat. Powered by the same MCP available to external agents like #Claude, Cursor and ChatGPT. Ask Violet about your data, summarize changes, generate charts and create reports. Chat is currently in Beta. If you want first dibs, shoot us a note or follow Violet Labs for updates.

The most exciting Violet feature we've shared to date. And it's not actually *in* Violet. You can now build entire apps, pulling in your live engineering data, with natural language prompts. The important part isn't the UI. It’s that models can now reason over structured engineering data across systems - powered by Violet's ontology and MCP. In about 20 minutes, our CTO Mike built this Siemens Teamcenter x First Resonance BOM Compare tool in #Claude. Sharing the Violet Labs how-to link in the comments.