{"passport":{"unfragile":{"@version":"1.0","version":"2026-05","artifact":{"id":"tool_triv-ai","slug":"triv-ai","name":"Triv AI","type":"webapp","url":"https://driv-ai.vercel.app","page_url":"https://unfragile.ai/triv-ai","categories":["app-builders"],"tags":[],"pricing":{"model":"free","free":true,"starting_price":null},"status":"active","verified":false},"capabilities":[{"id":"tool_triv-ai__cap_0","uri":"capability://planning.reasoning.adaptive.personalized.learning.path.generation","name":"adaptive-personalized-learning-path-generation","description":"Generates individualized learning sequences that adapt to detected knowledge gaps through real-time performance monitoring. The system tracks user responses to driving theory questions, identifies weak conceptual areas, and dynamically reorders or emphasizes curriculum modules to address deficiencies before progression. Implementation approach uses performance metrics (answer accuracy, response patterns, time-to-answer) to trigger curriculum branch selection, though specific ML model architecture (LLM-based, rule-based, or fine-tuned) is undocumented.","intents":["I want a driving theory course that doesn't force me through material I already know","I need the system to identify which driving concepts I'm struggling with and focus on those","I want my learning path to adjust based on my actual performance, not a fixed curriculum"],"best_for":["Adult learners with variable prior knowledge preparing for written driving exams","High school students who need flexible pacing around other commitments","Learners with identified weak areas in specific driving domains (e.g., traffic laws vs. vehicle mechanics)"],"limitations":["No documentation of how 'personalization' algorithm weights different performance signals — unclear if recency bias, accuracy threshold, or other metrics drive path changes","Adaptation scope limited to theory content only; no practical driving scenario integration mentioned","Unknown whether system can detect and adapt to learning style preferences (visual, kinesthetic, etc.) or only content mastery","No evidence of A/B testing or validation that adaptive paths improve learning outcomes vs. linear curriculum"],"requires":["Active internet connection for real-time session tracking","User account with persistent session storage (mechanism unknown)","Minimum engagement: completion of initial assessment or first lesson to establish baseline"],"input_types":["user responses to multiple-choice or short-answer driving theory questions","implicit signals: time spent per question, answer revision patterns, lesson completion timestamps"],"output_types":["ordered curriculum sequence (next lesson/module recommendation)","performance dashboard showing identified weak areas","progress metrics (% mastery by topic)"],"categories":["planning-reasoning","memory-knowledge"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"tool_triv-ai__cap_1","uri":"capability://text.generation.language.ai.driven.conversational.coaching.via.chatbot","name":"ai-driven-conversational-coaching-via-chatbot","description":"Delivers driving theory instruction and feedback through a conversational chatbot interface rather than traditional multiple-choice question banks. Users interact with an AI coach (implementation model unspecified: could be LLM-based like GPT/Claude, or rule-based dialogue system) that explains concepts, answers follow-up questions, and provides corrective feedback on user understanding. The chatbot maintains context within a session to enable multi-turn dialogue about driving scenarios and regulations.","intents":["I want to ask follow-up questions about driving concepts without being locked into multiple-choice answers","I prefer conversational learning over rigid question-and-answer formats","I need explanations of why certain driving rules exist, not just memorization of rules"],"best_for":["Learners who find traditional test-prep question banks intimidating or boring","Users seeking conceptual understanding rather than rote memorization for exam passing","Non-native speakers who benefit from conversational context over isolated questions"],"limitations":["No specification of which AI model powers the chatbot — unclear if responses are generated by fine-tuned LLM, prompt-engineered GPT, or rule-based dialogue engine","Unknown context window size — unclear if chatbot can maintain coherent multi-turn conversations over 10+ exchanges or resets context frequently","No mention of hallucination prevention or fact-checking mechanisms — risk of chatbot providing incorrect driving regulations","Scope limited to theoretical knowledge; chatbot cannot simulate real-world driving scenarios or provide feedback on actual vehicle operation","No documentation of response latency or whether 'real-time feedback' claim is quantified (e.g., <2 second response time)"],"requires":["Active internet connection for API calls to chatbot backend","User account with session state management","Web browser with JavaScript enabled for chatbot UI rendering"],"input_types":["natural language text queries about driving theory, regulations, vehicle mechanics","follow-up questions and clarification requests in conversational format"],"output_types":["natural language explanations of driving concepts","corrective feedback on user misconceptions","links or references to official driving regulations (if implemented)"],"categories":["text-generation-language","planning-reasoning"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"tool_triv-ai__cap_10","uri":"capability://text.generation.language.real.time.feedback.generation.on.user.responses","name":"real-time-feedback-generation-on-user-responses","description":"Generates immediate corrective feedback on user answers to driving theory questions and simulation decisions. The system evaluates user responses against correct answers/safe driving practices and provides explanations of why answers are correct/incorrect. Feedback is delivered via chatbot (natural language explanations) or structured messages (e.g., 'Incorrect: You should brake, not accelerate, when a pedestrian crosses'). Implementation approach (rule-based evaluation vs. LLM-generated explanations) is undocumented. Latency and quality of feedback are unspecified.","intents":["I want immediate feedback on whether my answer is correct","I want to understand WHY my answer was wrong, not just be told it's wrong","I want explanations that help me remember the correct answer for the exam"],"best_for":["Learners who benefit from immediate corrective feedback","Users seeking conceptual understanding over rote memorization","Learners with anxiety who want supportive, explanatory feedback"],"limitations":["Unknown latency of feedback delivery — 'real-time' is unquantified (could be <100ms or 5+ seconds)","No specification of feedback quality or whether explanations are accurate/helpful","Risk of LLM-generated feedback containing hallucinations or incorrect driving information","Unknown if feedback is personalized (e.g., addressing specific misconceptions) or generic templated responses","Unclear if feedback includes references to official regulations or just explanations","No mention of feedback quality assurance or human review of chatbot-generated explanations","Unknown if feedback adapts based on user learning style or prior performance"],"requires":["User submission of answer/response to question or simulation","Backend system with answer key and feedback templates/generation logic","Active internet connection for feedback delivery"],"input_types":["user answer to multiple-choice question","user decision in simulation scenario","user follow-up question about feedback"],"output_types":["correctness assessment (correct/incorrect/partially correct)","explanation of correct answer","reference to relevant driving regulation (if applicable)","encouragement or guidance for improvement"],"categories":["text-generation-language","planning-reasoning"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"tool_triv-ai__cap_2","uri":"capability://automation.workflow.interactive.driving.simulations.execution","name":"interactive-driving-simulations-execution","description":"Provides interactive simulations of driving scenarios to reinforce theoretical knowledge through practical application. The product claims 'interactive simulations' but provides no technical details on implementation (2D/3D graphics, physics engine, browser-based vs. external app, rule-based vs. ML-driven scenario generation). Simulations presumably present driving situations (e.g., 'traffic light turns red, pedestrian crossing ahead') and evaluate user decision-making against driving rules.","intents":["I want to practice decision-making in realistic driving scenarios without being in an actual car","I need to see how driving rules apply in context, not just memorize isolated regulations","I want feedback on whether my simulated driving decisions were safe and legal"],"best_for":["Learners preparing for written exams who want scenario-based reinforcement","Users seeking supplementary practice between in-person driving lessons","Learners with anxiety about real driving who want low-stakes scenario practice"],"limitations":["Critical unknown: no specification of simulation fidelity (2D diagrams vs. 3D graphics vs. video-based scenarios) — unclear if 'interactive' means user controls vehicle or just makes decision choices","No mention of physics simulation or whether vehicle behavior follows realistic dynamics or simplified rules","Scope appears limited to decision-making scenarios; no indication of full driving simulation (steering, acceleration, braking control)","Unknown whether scenarios are procedurally generated, templated, or pre-recorded — affects replay value and scenario diversity","No documentation of how feedback is generated or whether it aligns with official driving standards across different regions (US DMV vs. UK DVLA vs. other jurisdictions)","Unclear if simulations include edge cases (weather, night driving, highway merging) or only basic scenarios"],"requires":["Web browser with WebGL or similar graphics support (if 3D simulations)","Active internet connection for scenario delivery and feedback generation","Minimum screen size/resolution not specified","Completion of prior theory lessons (assumed prerequisite)"],"input_types":["user decisions in response to simulated driving scenarios (e.g., 'turn left', 'brake', 'proceed')","implicit signals: decision speed, confidence level (if UI captures it)"],"output_types":["scenario outcome (success/failure/unsafe decision)","feedback explaining why decision was correct/incorrect","performance metrics (% correct decisions, safety score)"],"categories":["automation-workflow","planning-reasoning"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"tool_triv-ai__cap_3","uri":"capability://text.generation.language.multi.language.content.delivery","name":"multi-language-content-delivery","description":"Delivers driving education content in multiple languages to serve non-English-speaking learners. Implementation approach is undocumented — unclear whether this is UI-only localization (buttons/menus translated) or full content translation (all driving theory, chatbot responses, simulation scenarios translated). Scope of language support and translation quality assurance mechanisms are not specified.","intents":["I want to learn driving theory in my native language, not English","I need driving regulations explained in a language I'm fluent in","I want to prepare for driving exams in my country using content in my language"],"best_for":["Non-English-speaking immigrants preparing for driving exams in English-speaking countries","International learners in non-English regions (Europe, Asia, etc.)","Multilingual households where different family members need different language versions"],"limitations":["No specification of which languages are supported — 'multi-language' could mean 2 languages or 50+","Unknown whether translation is human-curated or machine-translated — critical for accuracy of driving regulations","Unclear if all features (chatbot, simulations, feedback) are translated or only static content","No mention of regional driving standard variations — e.g., does Spanish version cover Spain's traffic laws or Latin America's?","Unknown if language selection is per-user or per-session — affects multi-user household usability","No documentation of translation update frequency if driving regulations change"],"requires":["User account with language preference setting","Backend support for language-specific content delivery (mechanism unknown)"],"input_types":["user language preference selection at signup or in settings"],"output_types":["all UI text in selected language","course content in selected language","chatbot responses in selected language (if implemented)"],"categories":["text-generation-language"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"tool_triv-ai__cap_4","uri":"capability://data.processing.analysis.progress.tracking.and.performance.analytics","name":"progress-tracking-and-performance-analytics","description":"Monitors user progress through the curriculum and generates performance analytics showing mastery levels by topic, completion rates, and weak areas. The system persists user state across sessions (mechanism unknown: likely database-backed user accounts) and aggregates performance signals (question accuracy, time-to-completion, simulation scores) into dashboards and reports. Enables users to resume learning from last checkpoint and track improvement over time.","intents":["I want to see which driving topics I've mastered and which I need to review","I need to track my progress toward completing the course and getting my 'mini license'","I want historical data on my performance to identify trends (e.g., improving or declining accuracy)"],"best_for":["Self-directed learners who benefit from quantified progress metrics","Users preparing for time-bound exams who need to track readiness","Parents/guardians monitoring teen driver education progress"],"limitations":["Unknown what metrics are tracked — could be basic (% completion) or detailed (response time, confidence levels, learning curve analysis)","No specification of dashboard UI or data visualization approach","Unclear if analytics are real-time or batch-processed — affects whether users see immediate feedback","Unknown if performance data is exportable or locked within platform","No mention of comparative analytics (e.g., 'you're in top 10% of learners') or benchmarking","Unclear if system detects learning plateaus or recommends intervention (e.g., 'take a break, review topic X')","Unknown data retention policy — what happens to progress data if subscription lapses?"],"requires":["User account with persistent session storage","Active internet connection to sync progress data","Minimum engagement: completion of at least one lesson to generate analytics"],"input_types":["implicit: user performance data (question responses, completion timestamps, simulation scores)","explicit: user review/retry actions"],"output_types":["progress dashboard showing % completion by topic","performance metrics (accuracy %, weak areas, strong areas)","historical charts showing progress over time","readiness assessment or 'mini license' eligibility status"],"categories":["data-processing-analysis","memory-knowledge"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"tool_triv-ai__cap_5","uri":"capability://automation.workflow.mini.driving.license.credential.issuance","name":"mini-driving-license-credential-issuance","description":"Issues a 'mini driving license' credential upon course completion as a gamification/motivation mechanism. The credential is explicitly NOT a legal driving license and has no jurisdictional recognition — it functions as a completion certificate or badge. Implementation approach (digital certificate, PDF download, blockchain-backed, shareable credential) is undocumented. Unclear whether credential is issued once per user or can be earned multiple times, and whether it includes metadata (completion date, topics mastered, score).","intents":["I want a tangible credential to show I've completed the driving theory course","I want to share proof of course completion with parents, employers, or driving schools","I want gamification/motivation to complete the full course"],"best_for":["High school students seeking completion proof for school records","Young drivers wanting to demonstrate theory knowledge to parents","Learners motivated by badges/credentials for course completion"],"limitations":["CRITICAL: 'Mini driving license' has ZERO legal standing — cannot be used as proof of driving eligibility in any jurisdiction","Unknown whether credential is recognized by driving schools, insurance companies, or employers","Unclear if credential includes performance data (score, topics mastered) or is generic completion badge","Unknown if credential is revoked if subscription lapses or if it's permanent","No mention of credential verification mechanism — unclear if third parties can validate authenticity","Scope limited to course completion; does not indicate readiness for official driving exams or practical driving tests","Unknown if credential can be revoked for misconduct or if it's non-revocable"],"requires":["Course completion (all modules/lessons finished)","Active subscription at time of issuance (assumed)","User account with credential storage/delivery mechanism"],"input_types":["implicit: course completion status and performance data"],"output_types":["digital credential (format unknown: PDF, image, blockchain token, etc.)","shareable link or downloadable file","metadata: completion date, user name, topics covered (if included)"],"categories":["automation-workflow"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"tool_triv-ai__cap_6","uri":"capability://automation.workflow.24.7.asynchronous.learning.access","name":"24-7-asynchronous-learning-access","description":"Enables learners to access course content, chatbot coaching, and simulations at any time without instructor availability constraints. The platform operates as a fully asynchronous, self-paced system with no live instructor sessions or scheduled class times. Users can start/pause/resume lessons independently, and the chatbot provides on-demand responses without human instructor involvement. Implementation relies on persistent backend infrastructure (database, API servers) to serve content and maintain session state across time zones and devices.","intents":["I want to study driving theory at midnight or early morning when I have free time","I need flexibility to fit driving education around my work/school schedule","I want to learn at my own pace without waiting for instructor availability"],"best_for":["Working adults with irregular schedules","High school students balancing multiple commitments","Learners in time zones with limited local driving school availability","Self-directed learners who prefer independent pacing"],"limitations":["No live instructor feedback — all coaching is AI-driven (chatbot) or pre-recorded","Unknown uptime SLA or availability guarantees — no mention of service level agreements","No mention of offline mode — requires continuous internet connectivity","Unclear if backend infrastructure can handle peak load (e.g., evening/weekend usage spikes)","No mention of data backup or disaster recovery — risk of progress loss if platform goes down","Asynchronous model may not suit learners who benefit from real-time instructor interaction or group learning"],"requires":["Active internet connection (no offline capability mentioned)","User account with session persistence across devices","Web browser with JavaScript enabled","Minimum uptime: platform must be operational (no SLA specified)"],"input_types":["user lesson start/pause/resume actions","chatbot queries at any time","simulation scenario responses"],"output_types":["course content delivered on-demand","chatbot responses within unspecified latency","progress updates persisted to user account"],"categories":["automation-workflow"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"tool_triv-ai__cap_7","uri":"capability://automation.workflow.subscription.based.access.control.and.billing","name":"subscription-based-access-control-and-billing","description":"Implements a subscription payment model with tiered pricing ($7 early-bird one-time, $20/month standard) to gate access to all platform features. The system manages user authentication, subscription state (active/expired/cancelled), and feature access based on payment status. Upon subscription expiration, access to lessons, chatbot, simulations, and progress data is revoked (mechanism for data retention post-expiration unknown). Billing is handled via unspecified payment processor (Stripe, PayPal, etc.) with no documented refund policy or cancellation terms.","intents":["I want to pay once for lifetime access to driving education content","I want to pay monthly and cancel anytime if the platform doesn't meet my needs","I want to know exactly what I'm paying for and when my subscription renews"],"best_for":["Budget-conscious learners seeking affordable driving education (vs. $2,000 traditional schools)","Users wanting short-term access (1-2 months) to prepare for specific exams","Learners in regions with limited driving school options"],"limitations":["CRITICAL: No freemium model — all features require payment; no free trial mentioned","Early-bird pricing ($7) limited to first 1,000 users; unclear if this offer is still active (website shows '0 sign-ups')","Unknown course completion timeline — unclear if $20/month is sufficient for 6-month course or requires 6+ months of subscription","No documented refund policy — unclear if users can get refunds if unsatisfied","No mention of family/household plans — each user requires separate subscription","Unknown what happens to progress data if subscription lapses — data may be deleted or inaccessible","No mention of payment method flexibility (credit card only? PayPal? Crypto?)","Unclear if subscription auto-renews or requires manual renewal","No documented cancellation process or ease of cancellation"],"requires":["Valid payment method (credit card, PayPal, etc. — not specified)","User account creation with email verification","Active subscription status to access any platform features"],"input_types":["payment method information (card details, PayPal account, etc.)","subscription tier selection (early-bird vs. standard)","billing address and personal information"],"output_types":["subscription confirmation and receipt","access token/session enabling feature access","billing history and renewal reminders","cancellation confirmation (if applicable)"],"categories":["automation-workflow"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"tool_triv-ai__cap_8","uri":"capability://automation.workflow.user.account.management.and.authentication","name":"user-account-management-and-authentication","description":"Manages user registration, login, password reset, and account settings. Users create accounts with email/password (or unspecified social login options) to access personalized learning paths and progress tracking. The system maintains user identity across sessions and devices, enabling seamless resume of interrupted lessons. Account management includes profile settings (language preference, learning goals) and subscription status visibility. Implementation uses standard web authentication patterns (likely JWT tokens or session cookies) with unspecified password security standards.","intents":["I want to create an account and start learning immediately","I want to access my learning progress from different devices (phone, laptop, tablet)","I want to reset my password if I forget it"],"best_for":["Individual learners preparing for driving exams","Users with multiple devices who want seamless cross-device access","Learners who value privacy and account security"],"limitations":["No documentation of password security standards (e.g., bcrypt hashing, minimum complexity requirements)","Unknown if two-factor authentication (2FA) is supported","No mention of social login options (Google, Apple, Facebook) — email/password only assumed","Unclear if account deletion is supported or if user data is retained indefinitely","No documented data privacy policy or GDPR compliance statement visible on homepage","Unknown if accounts can be shared (e.g., family household) or if each user requires separate subscription","No mention of parental controls for teen learners","Unclear if account recovery is possible if email is compromised"],"requires":["Valid email address for account creation and password reset","Password meeting unspecified complexity requirements","Internet connection for signup and login","Web browser with cookies/local storage enabled"],"input_types":["email address and password for signup/login","password reset token (via email)","profile settings (language, learning goals, etc.)"],"output_types":["account confirmation email","session token/cookie enabling authenticated access","user profile dashboard with settings","password reset confirmation"],"categories":["automation-workflow","safety-moderation"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"tool_triv-ai__cap_9","uri":"capability://text.generation.language.region.agnostic.driving.regulation.content.delivery","name":"region-agnostic-driving-regulation-content-delivery","description":"Delivers driving education content covering traffic laws and regulations, but with unknown regional scope and accuracy. The product claims to prepare users for 'driving license exams' but provides no specification of which jurisdictions' standards are covered (US DMV, UK DVLA, EU standards, etc.). Content accuracy and alignment with official driving authority regulations are undocumented. Unclear whether content is region-specific (e.g., separate US vs. UK curriculum) or generic 'universal' driving principles.","intents":["I want to learn driving regulations for my country/state","I want to prepare for my official driving license exam","I want to understand traffic laws specific to where I'll be driving"],"best_for":["Learners preparing for official driving exams in supported regions","Users seeking theory knowledge aligned with local traffic laws","Immigrants preparing for driving exams in new countries"],"limitations":["CRITICAL: No specification of which regions/jurisdictions are covered — unclear if US-only, global, or selective","No documentation of content accuracy or alignment with official DMV/DVLA/local standards","Unknown if content is reviewed/updated when regulations change","Unclear if regional variations are handled (e.g., right-hand vs. left-hand driving, speed limit differences)","No mention of third-party validation or regulatory authority endorsement","Unknown if content covers edge cases (commercial licenses, motorcycle endorsements, etc.) or basic passenger vehicle only","Risk of content being outdated or inaccurate, leading to exam failure or unsafe driving practices"],"requires":["User location/region selection at signup (mechanism unknown)","Internet connection to access region-specific content","Completion of prior lessons to access advanced topics"],"input_types":["user region/jurisdiction selection","implicit: user performance on region-specific questions"],"output_types":["region-specific driving theory content","practice questions aligned with local exam standards","feedback on answers relative to local regulations"],"categories":["text-generation-language","data-processing-analysis"],"confidence":0.5,"matches":0,"success_rate":0}],"trust":{"score":40,"verified":false,"data_access_risk":"high","permissions":["Active internet connection for real-time session tracking","User account with persistent session storage (mechanism unknown)","Minimum engagement: completion of initial assessment or first lesson to establish baseline","Active internet connection for API calls to chatbot backend","User account with session state management","Web browser with JavaScript enabled for chatbot UI rendering","User submission of answer/response to question or simulation","Backend system with answer key and feedback templates/generation logic","Active internet connection for feedback delivery","Web browser with WebGL or similar graphics support (if 3D simulations)"],"failure_modes":["No documentation of how 'personalization' algorithm weights different performance signals — unclear if recency bias, accuracy threshold, or other metrics drive path changes","Adaptation scope limited to theory content only; no practical driving scenario integration mentioned","Unknown whether system can detect and adapt to learning style preferences (visual, kinesthetic, etc.) or only content mastery","No evidence of A/B testing or validation that adaptive paths improve learning outcomes vs. linear curriculum","No specification of which AI model powers the chatbot — unclear if responses are generated by fine-tuned LLM, prompt-engineered GPT, or rule-based dialogue engine","Unknown context window size — unclear if chatbot can maintain coherent multi-turn conversations over 10+ exchanges or resets context frequently","No mention of hallucination prevention or fact-checking mechanisms — risk of chatbot providing incorrect driving regulations","Scope limited to theoretical knowledge; chatbot cannot simulate real-world driving scenarios or provide feedback on actual vehicle operation","No documentation of response latency or whether 'real-time feedback' claim is quantified (e.g., <2 second response time)","Unknown latency of feedback delivery — 'real-time' is unquantified (could be <100ms or 5+ seconds)","builder identity is not verified yet","no observed match outcomes yet"],"rank_breakdown":{"adoption":0.31666666666666665,"quality":0.72,"ecosystem":0.15000000000000002,"match_graph":0.25,"freshness":0.75,"weights":{"adoption":0.25,"quality":0.25,"ecosystem":0.1,"match_graph":0.35,"freshness":0.05}},"observed_outcomes":{"matches":0,"success_rate":0,"avg_confidence":0,"top_intents":[],"last_matched_at":null},"maintenance":{"status":"active","updated_at":"2026-05-24T12:16:33.648Z","last_scraped_at":"2026-04-05T13:23:42.559Z","last_commit":null},"community":{"stars":null,"forks":null,"weekly_downloads":null,"model_downloads":null,"model_likes":null}},"distribution":{"claim_url":"https://unfragile.ai/submit?claim=triv-ai","compare_url":"https://unfragile.ai/compare?artifact=triv-ai"}},"signature":"J2BNqxRhxXk6PKwW8VfSuNX2xcDbO3qefdR3Z/6m27wa1KBi/UBPJ+Phrzj1UC1y1tttPB5rzKUNwKo9YqXgCw==","signedAt":"2026-06-20T22:48:23.949Z","signedBy":"unfragile.ai","version":1},"_links":{"self":"https://unfragile.ai/api/v1/passport/triv-ai","artifact":"https://unfragile.ai/triv-ai","verify":"https://unfragile.ai/api/v1/verify?slug=triv-ai","publicKey":"https://unfragile.ai/api/v1/trust-passport-public-key","spec":"https://unfragile.ai/trust","schema":"https://unfragile.ai/schema.json","docs":"https://unfragile.ai/docs"}}