{"passport":{"unfragile":{"@version":"1.0","version":"2026-05","artifact":{"id":"tool_breadcrumb-ai","slug":"breadcrumb-ai","name":"Breadcrumb.ai","type":"product","url":"https://www.breadcrumb.ai","page_url":"https://unfragile.ai/breadcrumb-ai","categories":["data-analysis"],"tags":[],"pricing":{"model":"freemium","free":true,"starting_price":null},"status":"active","verified":false},"capabilities":[{"id":"tool_breadcrumb-ai__cap_0","uri":"capability://data.processing.analysis.real.time.data.ingestion.and.transformation.pipeline","name":"real-time data ingestion and transformation pipeline","description":"Breadcrumb.ai ingests raw data from multiple sources (marketing platforms, analytics tools, databases) and applies automated transformation logic to normalize, deduplicate, and enrich datasets in real-time. The system likely uses event-streaming architecture (Kafka-like patterns) or webhook-based connectors to capture data changes and apply transformation rules without batch delays, enabling sub-minute latency for dashboard updates.","intents":["I need to connect my marketing data sources (Google Analytics, HubSpot, Salesforce) and have them automatically normalized into a unified schema without manual ETL","I want raw event data transformed and aggregated in real-time so my dashboards reflect current campaign performance, not yesterday's numbers","I need to deduplicate and clean messy marketing data (duplicate leads, inconsistent naming) automatically as it flows in"],"best_for":["Marketing operations teams managing multiple disconnected data sources","Demand generation managers who need fresh data for daily decision-making","Non-technical marketers who lack SQL/Python skills for custom ETL"],"limitations":["Transformation rules are likely limited to pre-built connectors — custom transformations may require API access or manual configuration","Real-time processing adds latency overhead; complex aggregations may still require eventual consistency","Data quality depends entirely on source system quality — garbage in, garbage out applies even with automated transformation"],"requires":["API credentials for connected data sources (Google Analytics, HubSpot, Salesforce, etc.)","Network connectivity to source systems for webhook/polling-based ingestion","Minimum data volume threshold (unknown, likely 1000+ events/day for meaningful real-time processing)"],"input_types":["structured data (CSV, JSON from APIs)","event streams (webhook payloads)","database connections (SQL queries)","marketing platform APIs (HubSpot, Salesforce, Google Analytics)"],"output_types":["normalized structured data (JSON, Parquet)","aggregated metrics (counts, sums, averages)","enriched records with computed fields"],"categories":["data-processing-analysis","automation-workflow"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"tool_breadcrumb-ai__cap_1","uri":"capability://text.generation.language.ai.driven.narrative.generation.from.metrics.and.trends","name":"ai-driven narrative generation from metrics and trends","description":"Breadcrumb.ai applies large language models to structured marketing metrics, time-series data, and statistical patterns to automatically generate human-readable narratives that explain what happened, why it matters, and what to do next. The system likely uses prompt engineering with metric context (deltas, anomalies, benchmarks) to produce coherent storytelling that translates raw numbers into actionable insights without requiring manual interpretation.","intents":["I want the system to automatically explain why my email open rate dropped 15% this week in plain English, not just show me the number","I need AI to identify which campaigns are underperforming and suggest why (audience fatigue, poor timing, creative issues) based on the data","I want daily narrative summaries of marketing performance emailed to executives who don't read dashboards"],"best_for":["Marketing executives and non-technical stakeholders who need insights without dashboard literacy","Demand generation teams who want automated anomaly detection and explanation","Organizations with high data literacy who want to validate AI-generated insights against domain knowledge"],"limitations":["AI narratives can oversimplify complex multivariate phenomena — correlation may be misattributed to causation","Generated insights are only as good as the underlying metrics; missing KPIs or poor data quality produces misleading narratives","No transparency into which metrics the AI prioritized or how confidence scores were calculated — black-box interpretation risk","Narratives may miss domain-specific context (e.g., seasonal patterns, competitive actions, external events) that humans would catch"],"requires":["Structured metric data with temporal dimension (timestamps, values, dimensions)","Baseline or benchmark data for anomaly detection (historical averages, targets)","LLM API access (OpenAI, Anthropic, or proprietary model) with sufficient rate limits"],"input_types":["time-series metrics (JSON with timestamp, value, dimension)","aggregated KPIs (conversion rate, revenue, engagement)","categorical dimensions (campaign, channel, audience segment)"],"output_types":["natural language narrative (plain text, markdown)","structured insights (JSON with key findings, anomalies, recommendations)","email-ready summaries (HTML formatted)"],"categories":["text-generation-language","data-processing-analysis"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"tool_breadcrumb-ai__cap_10","uri":"capability://planning.reasoning.predictive.trend.analysis.and.forecasting","name":"predictive trend analysis and forecasting","description":"Breadcrumb.ai applies time-series forecasting models (ARIMA, exponential smoothing, or machine learning-based) to historical metric data to predict future values and trends. The system likely generates forecasts with confidence intervals and uses them to contextualize current performance (e.g., 'conversion rate is tracking 5% below forecast') and alert users to deviations from expected trajectory.","intents":["I want to forecast next month's revenue based on current pipeline and historical conversion rates","I need to predict when we'll hit our quarterly lead generation target based on current velocity","I want to see if current campaign performance is on track to meet goals or if we need to adjust strategy"],"best_for":["Demand generation managers who need to forecast lead volume and pipeline","Marketing leaders who need to predict revenue impact of campaigns","Teams that want to proactively identify if they're off-track vs goals"],"limitations":["Forecasts are only as good as historical data — new products, markets, or campaigns have no baseline for prediction","Forecasting models assume historical patterns continue — can't account for structural changes (new competitor, market shift, product launch)","Confidence intervals may be wide for volatile metrics, reducing actionability","Forecasts don't account for planned business actions (budget increases, new campaigns) — require manual adjustment"],"requires":["Minimum 3-6 months of historical metric data for model training","Consistent metric definition over time (can't change how metric is calculated mid-stream)","Reasonable data volume and frequency (daily or higher for meaningful forecasts)"],"input_types":["historical time-series metric data (timestamp, value)","forecast horizon (how many periods ahead to predict)","confidence level (e.g., 80%, 95%)"],"output_types":["forecast values with confidence intervals (point estimate, lower bound, upper bound)","forecast accuracy metrics (MAPE, RMSE from historical backtesting)","trend analysis (increasing, decreasing, stable)"],"categories":["planning-reasoning","data-processing-analysis"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"tool_breadcrumb-ai__cap_2","uri":"capability://data.processing.analysis.real.time.interactive.dashboard.with.metric.visualization","name":"real-time interactive dashboard with metric visualization","description":"Breadcrumb.ai renders live dashboards that update as new data arrives, displaying metrics, trends, and KPIs with interactive filtering and drill-down capabilities. The system likely uses a client-side charting library (D3.js, Plotly, or similar) with WebSocket/Server-Sent Events for real-time updates, allowing users to explore data without page refreshes while maintaining performance at scale.","intents":["I need a live dashboard showing campaign performance metrics that updates every minute without manual refresh","I want to filter dashboard data by campaign, channel, or audience segment and see results update instantly","I need to drill down from high-level KPIs (total revenue) to granular details (revenue by campaign by day) without leaving the dashboard"],"best_for":["Marketing operations teams monitoring live campaign performance","Demand generation managers who need real-time visibility into lead generation metrics","Teams running time-sensitive campaigns (product launches, flash sales) where stale data is costly"],"limitations":["Real-time updates require persistent WebSocket connections — may not scale to thousands of concurrent users without infrastructure investment","Complex aggregations or large datasets may cause dashboard lag or timeout; pre-aggregation required for performance","Limited customization compared to enterprise BI tools — dashboard layouts and visualizations likely templated rather than fully flexible"],"requires":["Modern web browser with WebSocket support (Chrome, Firefox, Safari, Edge)","Stable internet connection for real-time streaming","Data sources configured and actively ingesting (see real-time data ingestion capability)"],"input_types":["structured metrics (JSON time-series)","categorical dimensions (campaign, channel, audience)","user filter selections (date range, segment, metric)"],"output_types":["interactive visualizations (line charts, bar charts, heatmaps)","filtered metric tables (CSV export capable)","drill-down navigation (click to expand detail)"],"categories":["data-processing-analysis","automation-workflow"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"tool_breadcrumb-ai__cap_3","uri":"capability://tool.use.integration.multi.source.data.connector.framework.with.pre.built.integrations","name":"multi-source data connector framework with pre-built integrations","description":"Breadcrumb.ai provides a connector library that abstracts authentication, API pagination, and schema mapping for popular marketing and analytics platforms (Google Analytics, HubSpot, Salesforce, Facebook Ads, LinkedIn Ads, etc.). Each connector likely implements a standardized interface that handles OAuth/API key management, incremental syncs, and field mapping to a common schema, reducing integration effort from weeks to minutes.","intents":["I want to connect HubSpot and Google Analytics without writing custom API code or managing authentication myself","I need to sync data from multiple ad platforms (Facebook, LinkedIn, Google Ads) into a single unified view without manual exports","I want new data sources added to my dashboard without engineering involvement — just click 'add connector' and authenticate"],"best_for":["Marketing teams with limited engineering resources who need quick integrations","Organizations using standard marketing tech stacks (HubSpot, Salesforce, Google Analytics)","Teams who want to avoid building and maintaining custom API integrations"],"limitations":["Limited to pre-built connectors — custom data sources or legacy systems require API access or manual CSV uploads","Connector availability and freshness depends on Breadcrumb.ai's roadmap; gaps in coverage may block adoption","Schema mapping is likely opinionated — custom field transformations may not be supported without manual configuration","Rate limits on source APIs may cause sync delays during high-traffic periods"],"requires":["API credentials or OAuth tokens for each connected platform","Appropriate permissions in source systems (e.g., 'view analytics' in Google Analytics, 'read contacts' in HubSpot)","Network connectivity to source APIs"],"input_types":["OAuth authorization flows","API credentials (keys, tokens)","field mapping configuration (source field → target field)"],"output_types":["normalized data in unified schema","sync logs and error reports","field mapping metadata"],"categories":["tool-use-integration","data-processing-analysis"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"tool_breadcrumb-ai__cap_4","uri":"capability://data.processing.analysis.anomaly.detection.and.alerting.on.metric.deviations","name":"anomaly detection and alerting on metric deviations","description":"Breadcrumb.ai monitors metric time-series data and automatically detects statistical anomalies (sudden spikes, drops, or trend breaks) using statistical methods (z-score, isolation forest, or similar) or learned baselines. When anomalies are detected, the system generates alerts and narratives explaining the deviation, enabling teams to catch problems or opportunities without manual monitoring.","intents":["I want to be alerted immediately if my email open rate drops below normal, so I can investigate before the campaign ends","I need the system to flag unusual spikes in website traffic or conversion rate and explain what might have caused them","I want to set custom thresholds for each metric (e.g., alert if revenue drops >20% or engagement spikes >50%)"],"best_for":["Marketing operations teams who need early warning of campaign issues","Demand generation managers monitoring lead quality and volume in real-time","Teams running continuous campaigns where anomalies require immediate action"],"limitations":["Anomaly detection requires historical baseline data — new metrics or campaigns may not trigger alerts until sufficient history is collected","Statistical methods may produce false positives (alert fatigue) if thresholds are too sensitive, or miss real anomalies if too conservative","Explanations for anomalies are AI-generated and may be speculative — root cause analysis still requires human investigation","Seasonal patterns or known external events (holidays, product launches) may not be automatically excluded from anomaly detection"],"requires":["Minimum 2-4 weeks of historical metric data for baseline calculation","Configured alert channels (email, Slack, webhook)","Real-time data ingestion (see data ingestion capability)"],"input_types":["time-series metrics (timestamp, value, dimension)","alert threshold configuration (absolute value, percentage change, z-score)","baseline period specification (e.g., last 30 days)"],"output_types":["alert notifications (email, Slack, webhook payload)","anomaly metadata (detected value, baseline, deviation magnitude)","AI-generated explanation narrative"],"categories":["data-processing-analysis","planning-reasoning"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"tool_breadcrumb-ai__cap_5","uri":"capability://data.processing.analysis.metric.definition.and.custom.kpi.builder","name":"metric definition and custom kpi builder","description":"Breadcrumb.ai allows users to define custom metrics and KPIs by composing raw data fields with mathematical operations (sum, average, ratio, growth rate) and filters without writing SQL. The system likely uses a visual metric builder or formula language that translates user definitions into optimized queries, enabling non-technical marketers to create derived metrics and track them across dashboards and narratives.","intents":["I want to create a custom KPI like 'cost per qualified lead' by dividing ad spend by leads that match our ICP criteria","I need to define 'marketing-influenced revenue' as revenue from deals where marketing touched the account in the last 90 days","I want to track 'campaign efficiency' as (conversions / impressions) and compare it across channels without writing SQL"],"best_for":["Marketing teams with domain expertise but limited SQL/analytics skills","Organizations with custom KPIs that don't map to standard metrics","Teams that need to iterate on metric definitions frequently as strategy evolves"],"limitations":["Visual metric builders are limited to simple operations — complex statistical calculations or machine learning models may require SQL access","Metric definitions are likely stored in proprietary format — exporting or migrating to other BI tools may be difficult","Performance depends on underlying data volume — complex metrics on large datasets may cause dashboard lag","No version control or audit trail for metric definition changes — hard to track when/why metrics were modified"],"requires":["Access to raw data fields and their data types","Understanding of metric composition (which fields to combine, what operations to apply)","Sufficient data volume to calculate metrics without excessive latency"],"input_types":["raw data fields (from connected sources)","mathematical operators (+, -, *, /, %, growth rate)","filter conditions (date range, dimension values)","aggregation functions (sum, average, count, distinct count)"],"output_types":["metric definition (formula, filters, aggregation)","calculated metric values (time-series)","metric metadata (name, description, last updated)"],"categories":["data-processing-analysis","planning-reasoning"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"tool_breadcrumb-ai__cap_6","uri":"capability://data.processing.analysis.comparative.analysis.and.benchmarking.across.dimensions","name":"comparative analysis and benchmarking across dimensions","description":"Breadcrumb.ai enables users to compare metrics across dimensions (campaigns, channels, audiences, time periods) and automatically generates insights about relative performance, winners/losers, and trends. The system likely uses statistical comparison methods (t-tests, effect sizes) and visualization techniques (side-by-side charts, ranking tables) to surface meaningful differences and contextualize performance within the broader dataset.","intents":["I want to compare email campaign performance across audience segments and see which segment has the highest engagement","I need to benchmark this month's conversion rate against last month and last year to understand if we're improving","I want to rank all my ad campaigns by ROI and see which channels are outperforming expectations"],"best_for":["Marketing managers who need to allocate budget across channels or campaigns based on performance","Demand generation teams comparing lead quality across sources","Organizations with multiple campaigns or segments that need performance ranking"],"limitations":["Comparisons are limited to dimensions present in the data — can't compare across external benchmarks (industry averages, competitor data) without manual import","Statistical significance testing may not be applied — differences may be due to random variation rather than real performance gaps","Comparative narratives may not account for confounding variables (e.g., audience size differences, seasonal effects) that explain performance gaps","Large numbers of dimensions may produce overwhelming comparison matrices without intelligent filtering or summarization"],"requires":["Data with multiple dimension values (e.g., multiple campaigns, channels, audience segments)","Sufficient sample size per dimension for meaningful comparison (minimum 30-100 events per group)","Metric definitions that are comparable across dimensions (e.g., same conversion definition for all channels)"],"input_types":["metric values by dimension (campaign, channel, audience, time period)","dimension filters (which dimensions to compare)","comparison type (side-by-side, ranking, trend)"],"output_types":["comparative visualizations (bar charts, ranking tables, trend overlays)","statistical summaries (mean, median, standard deviation per dimension)","narrative insights (which dimension is winning, by how much, confidence level)"],"categories":["data-processing-analysis","planning-reasoning"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"tool_breadcrumb-ai__cap_7","uri":"capability://automation.workflow.scheduled.report.generation.and.distribution","name":"scheduled report generation and distribution","description":"Breadcrumb.ai allows users to define report templates (combining dashboards, metrics, and AI-generated narratives) and schedule them to be generated and distributed automatically via email, Slack, or webhook. The system likely uses a scheduling engine (cron-like) to trigger report generation on a cadence, render the report with current data, and deliver it to specified recipients without manual intervention.","intents":["I want a daily email report showing yesterday's marketing performance, sent to my team every morning at 9 AM","I need a weekly summary of campaign performance with AI-generated insights posted to our Slack channel","I want to send a monthly executive dashboard to leadership automatically without manually creating it each month"],"best_for":["Marketing teams with recurring reporting needs (daily, weekly, monthly)","Organizations where executives expect consistent, timely reports without manual effort","Teams that want to reduce time spent on report creation and distribution"],"limitations":["Report templates are likely static — can't dynamically adjust content based on data anomalies or business context","Scheduled reports use data as of generation time — can't include real-time data if report is generated during off-hours","Email delivery may be unreliable or land in spam; Slack integration depends on workspace permissions","No built-in version control or audit trail for report changes — hard to track when/why report definitions were modified"],"requires":["Report template definition (which metrics, dashboards, narratives to include)","Schedule configuration (frequency, time, timezone)","Distribution channel configuration (email addresses, Slack channel, webhook URL)","Sufficient data freshness for scheduled time (e.g., if report runs at 6 AM, data must be available by then)"],"input_types":["report template (dashboard, metrics, narrative sections)","schedule (cron expression or UI-based frequency)","distribution targets (email, Slack, webhook)"],"output_types":["rendered report (HTML email, Slack message, PDF attachment)","delivery logs (sent, failed, bounced)","report metadata (generated timestamp, data freshness)"],"categories":["automation-workflow","text-generation-language"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"tool_breadcrumb-ai__cap_8","uri":"capability://data.processing.analysis.data.quality.monitoring.and.validation","name":"data quality monitoring and validation","description":"Breadcrumb.ai monitors incoming data for quality issues (missing values, outliers, schema violations, duplicate records) and flags problems before they corrupt dashboards or narratives. The system likely uses data profiling techniques (null rate analysis, cardinality checks, distribution analysis) and configurable validation rules to detect and alert on data quality degradation, enabling teams to catch source system issues early.","intents":["I want to be alerted if a data source stops sending data or starts sending null values for critical fields","I need to catch duplicate records or malformed data before they skew my metrics and dashboards","I want to validate that incoming data matches expected schema and value ranges (e.g., revenue should be positive)"],"best_for":["Data-driven marketing teams who depend on data quality for decision-making","Organizations with complex data pipelines where quality issues are common","Teams that want to prevent bad data from corrupting dashboards and narratives"],"limitations":["Data quality rules are likely pre-configured or require manual definition — can't automatically detect all types of quality issues","Quality monitoring adds latency to data ingestion — may delay dashboard updates if validation is synchronous","Alerts may produce false positives if rules are too strict, or miss real issues if too lenient","No automatic remediation — quality issues are flagged but require human investigation and source system fixes"],"requires":["Data schema definition (expected fields, data types, constraints)","Quality rule configuration (null rate thresholds, outlier detection, cardinality limits)","Alert channel configuration (email, Slack, webhook)"],"input_types":["incoming data records (JSON, CSV, database rows)","quality rule definitions (null rate, outlier, schema, cardinality)","baseline profiles (expected distributions, value ranges)"],"output_types":["quality reports (null rates, outlier counts, schema violations)","quality alerts (data quality issue detected, severity level)","remediation suggestions (likely causes, recommended fixes)"],"categories":["data-processing-analysis","safety-moderation"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"tool_breadcrumb-ai__cap_9","uri":"capability://automation.workflow.collaborative.insights.sharing.and.annotation","name":"collaborative insights sharing and annotation","description":"Breadcrumb.ai enables teams to share dashboards, reports, and AI-generated narratives with colleagues and add annotations, comments, or context to specific metrics or findings. The system likely uses role-based access control to manage who can view/edit dashboards, and provides commenting or annotation features to facilitate discussion and knowledge sharing without requiring email or separate tools.","intents":["I want to share a dashboard with my team and have them comment on specific metrics or insights","I need to annotate a metric with context (e.g., 'this spike was due to a promotional campaign') so others understand what happened","I want to control who can view or edit dashboards based on their role (viewer, editor, admin)"],"best_for":["Marketing teams that need to collaborate on data interpretation and decision-making","Organizations with distributed teams that need asynchronous collaboration on insights","Teams that want to build institutional knowledge about what metrics mean and why they changed"],"limitations":["Collaboration features are likely basic (comments, annotations) — no advanced features like version control, change tracking, or approval workflows","Access control is likely role-based — can't implement fine-grained permissions (e.g., 'can view this dashboard only for this date range')","Annotations are stored in Breadcrumb.ai — can't export or migrate them to other tools","No integration with external collaboration tools (Slack, Teams, Jira) for notifications or workflow integration"],"requires":["Team members with Breadcrumb.ai accounts","Configured access control (roles, permissions)","Shared dashboard or report URL"],"input_types":["dashboard or report to share","user email addresses or team names","access level (viewer, editor, admin)","annotation text and target metric"],"output_types":["shared dashboard URL with access control","annotation metadata (author, timestamp, text)","collaboration activity log (who viewed, commented, when)"],"categories":["automation-workflow","text-generation-language"],"confidence":0.5,"matches":0,"success_rate":0}],"trust":{"score":43,"verified":false,"data_access_risk":"high","permissions":["API credentials for connected data sources (Google Analytics, HubSpot, Salesforce, etc.)","Network connectivity to source systems for webhook/polling-based ingestion","Minimum data volume threshold (unknown, likely 1000+ events/day for meaningful real-time processing)","Structured metric data with temporal dimension (timestamps, values, dimensions)","Baseline or benchmark data for anomaly detection (historical averages, targets)","LLM API access (OpenAI, Anthropic, or proprietary model) with sufficient rate limits","Minimum 3-6 months of historical metric data for model training","Consistent metric definition over time (can't change how metric is calculated mid-stream)","Reasonable data volume and frequency (daily or higher for meaningful forecasts)","Modern web browser with WebSocket support (Chrome, Firefox, Safari, Edge)"],"failure_modes":["Transformation rules are likely limited to pre-built connectors — custom transformations may require API access or manual configuration","Real-time processing adds latency overhead; complex aggregations may still require eventual consistency","Data quality depends entirely on source system quality — garbage in, garbage out applies even with automated transformation","AI narratives can oversimplify complex multivariate phenomena — correlation may be misattributed to causation","Generated insights are only as good as the underlying metrics; missing KPIs or poor data quality produces misleading narratives","No transparency into which metrics the AI prioritized or how confidence scores were calculated — black-box interpretation risk","Narratives may miss domain-specific context (e.g., seasonal patterns, competitive actions, external events) that humans would catch","Forecasts are only as good as historical data — new products, markets, or campaigns have no baseline for prediction","Forecasting models assume historical patterns continue — can't account for structural changes (new competitor, market shift, product launch)","Confidence intervals may be wide for volatile metrics, reducing actionability","builder identity is not verified yet","no observed match outcomes yet"],"rank_breakdown":{"adoption":0.36666666666666664,"quality":0.78,"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":"2026-05-24T12:16:29.715Z","last_scraped_at":"2026-04-05T13:23:42.552Z","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=breadcrumb-ai","compare_url":"https://unfragile.ai/compare?artifact=breadcrumb-ai"}},"signature":"3717TdbFG6AiIUcUp8d3XP9wCt+aqkzACj/l+xgBPMVBt56wtwDrki/+oUlKHaSBdCNTyFv7CypMxvhBIxcWAw==","signedAt":"2026-06-20T21:19:49.736Z","signedBy":"unfragile.ai","version":1},"_links":{"self":"https://unfragile.ai/api/v1/passport/breadcrumb-ai","artifact":"https://unfragile.ai/breadcrumb-ai","verify":"https://unfragile.ai/api/v1/verify?slug=breadcrumb-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"}}