The ideas behind Lexicon.
Every product decision in Lexicon was driven by a specific founder pain point, a gap in the existing market, or a design principle we were not willing to compromise on.
A founder signed a vendor agreement and spent $31,000 fixing a clause they did not understand.
The indemnification clause required them to defend the vendor against any claim arising from use of the platform. When the vendor was sued by a third party over IP, the founder's startup was dragged into litigation they could not afford. Lexicon exists so this does not happen to the next founder.
Design decisions
We optimize for next actions, not summaries
Every other tool tells you what the contract says. We focus on what you should DO: negotiate this, remove this, add this, sign as-is. The output is decision support, not a transcript.
Severity labels instead of vague warnings
CRITICAL, WARN, and INFO give founders an instant priority stack. You immediately know what needs to be fixed before signing and what is just worth knowing.
Demo mode built for live presentations
Every hackathon pitch involves live demos. Demo mode lets you switch between three different risk levels - low, medium, high - instantly and reliably, without needing an internet connection or API key.
Plain English is not optional
Every warning includes three layers: what the clause says (detail), what it means for you specifically (whatItMeans), and what to do about it (whatToDo). No legal jargon without immediate translation.
BYO API key for real analysis
For founders who want to analyze their own real contracts during the demo, we support any OpenAI-compatible model via a user-supplied API key, stored locally in the browser.
Placeholder pricing that communicates the model
The pricing page is not functional - it is a communication tool. It shows judges and potential users that we understand how to monetize: per-seat plus usage, with team and enterprise tiers.
What the existing tools get wrong.
3 to 7 days turnaround, $300 to $800 per hour. Completely inaccessible for early-stage founders reviewing a routine vendor agreement. A $2,000 legal review on a $5,000 contract is not viable.
These tools generate and store documents. They do not analyze them. There is no risk detection, no warning system, no negotiation output. Founders still cannot tell if what they are signing is dangerous.
Unstructured output that varies every time. No severity ranking, no specific edits, no urgency structure. Requires significant prompting expertise from the founder to get useful output.