Build an Autonomous Trademark Conflict Monitor with Mastra AI
The high-profile trademark dispute between Jodii and Jodi365 highlights legal compliance risks faced by Indian D2C and consumer internet brands. In this challenge, you will implement an autonomous trademark conflict monitoring and IP intelligence agent using Mastra AI. The agent will ingest public registry filings, compute phonetic and visual similarity scores across trademark databases, perform LLM-based brand risk assessment, and trigger proactive notification workflows for brand legal teams.
What you are building
The core problem, expected build, and operating context for this challenge.
Build an IP risk monitoring workflow using Mastra AI TypeScript agents to compute trademark phonetic and visual similarity.
How work is evaluated
Evaluates phonetic similarity calculation and risk level assignment for brand name conflicts.
Challenge datasets
Public datasets, plus the private datasets you have access to.
How submissions are scored
These dimensions define what the evaluator checks and which criteria separate a passable run from a strong one.
high_conflict_detection_test
Ensures near-identical brand mark within the same class is flagged as HIGH conflict risk.
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
accuracy
Accuracy in distinguishing conflicting vs non-conflicting trademark pairs. • target: 0.95 • range: 0-1
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
What you should walk away with
Build Mastra AI TypeScript workflow steps with built-in memory and state management.
Implement phonetic similarity algorithms (Double Metaphone, Soundex) for Indic brand names.
Calculate Levenshtein and embedding distance between competing brand registrations.
Generate structured legal risk opinion summaries for trademark counsel.
Reference links and supporting material
Trademark database registry entries from Indian Intellectual Property Office filings with class categories.
Challenge setup
Evaluates phonetic similarity calculation and risk level assignment for brand name conflicts.
Challenge input
JSON with registered_mark, new_application_mark, class_category.
Mastra AI
Required TypeScript agent framework.
Evaluated output
JSON with phonetic_similarity, conflict_risk, legal_recommendation.
- Ensures near-identical brand mark within the same class is flagged as HIGH conflict risk.
- Accuracy in distinguishing conflicting vs non-conflicting trademark pairs. • target: 0.95 • range: 0-1
- Benchmark: Legal NLP Trademark Benchmark
- Accuracy target: 0.95
- 1 public reference case
- JavaScript execution harness
- javascript sandbox (unavailable on Versalist)
Summary shows the first task, up to three tools, and up to three evidence items. Open the technical recipe to view all configured items.
View technical recipe
Configured tools
- Mastra AI · Required
- Kubeflow · Optional
- Bito AI · Optional
- Mastra AI · Required
Evaluation contract
- high_conflict_detection_test · Weight 1
- accuracy · Weight 1
Metrics
- accuracy · Target 0.95
Benchmarks
- Legal NLP Trademark Benchmark
- general
Reference cases
- 1 public reference case
Workflow tasks
- Task 1
Input: JSON with registered_mark, new_application_mark, class_category.
Output: JSON with phonetic_similarity, conflict_risk, legal_recommendation.
Execution harness: JavaScript
Evaluation tools
- Mastra AI · Agent Framework
Recipe freeze marker
No recipe freeze time is recorded.