footprintjs

Automatically instruments backend execution paths to generate causal traces showing how data flows through functions, API calls, and decision points. Uses AST analysis and runtime instrumentation to capture the dependency graph between inputs, intermediate states, and outputs without requiring manual annotation. Traces are structured as directed acyclic graphs (DAGs) that can be serialized and replayed for debugging or audit purposes.

Extracts the evidence, conditions, and decision rules that led to a specific backend outcome, then generates human-readable narratives explaining the decision chain. Analyzes the causal trace to identify which inputs were actually used in the decision (vs. which were available but ignored), reconstructs the logical conditions that were evaluated, and produces structured evidence objects that can be presented to users or AI agents. Supports template-based narrative generation for different audiences (technical, business, regulatory).

Automatically generates Model Context Protocol (MCP) tool definitions from instrumented backend functions and API endpoints, creating structured schemas that describe inputs, outputs, side effects, and decision logic. Analyzes the causal traces and evidence extraction to infer tool semantics (e.g., 'this function filters users by criteria and returns a ranked list'), generates OpenAPI-compatible schemas with proper type definitions, and produces MCP tool manifests that AI agents can consume. Includes automatic documentation generation from code comments and inferred behavior.

Captures immutable state snapshots at each step of a causal trace, enabling developers to inspect the exact state of variables, function arguments, and return values at any point in the execution. Provides a queryable interface to jump to specific trace steps, inspect state diffs between consecutive steps, and replay execution from any checkpoint. Uses structural sharing and delta compression to minimize memory overhead while maintaining full state history.

Integrates with rule engines and decision tree systems to automatically instrument rule evaluation, capture which rules matched/failed, and visualize the decision tree structure with execution paths highlighted. Supports multiple rule engine formats (JSON-based rules, Drools-style syntax, custom DSLs) and generates interactive flowchart visualizations showing the decision path taken during execution. Includes rule conflict detection and coverage analysis to identify unreachable rules or conflicting conditions.

Provides state management for multi-step backend workflows and pipelines, automatically tracking state transitions, validating state changes against defined schemas, and enabling rollback to previous states. Integrates with causal tracing to record why state changed (which function triggered it, what conditions were met), and supports compensation logic for undoing operations in reverse order. Includes built-in support for saga patterns and distributed transaction coordination across service boundaries.

Automatically generates structured logs from causal traces, integrating with standard observability platforms (Datadog, New Relic, CloudWatch, ELK). Converts trace data into structured log entries with proper correlation IDs, trace IDs, and span hierarchies compatible with OpenTelemetry standards. Enables querying and filtering logs by decision evidence, rule matches, and state changes rather than just text search. Includes automatic sampling and aggregation for high-volume systems to reduce storage costs.

Automatically generates compliance and audit reports from causal traces, decision evidence, and state histories. Supports multiple report formats (PDF, HTML, JSON) and compliance frameworks (GDPR, HIPAA, SOX, Fair Lending). Includes data lineage tracking to show which personal data was used in decisions, automatic redaction of sensitive information, and proof of decision rationale for regulatory review. Generates attestation documents showing that decisions were made according to defined rules and policies.