footprintjs
core engineThe flowchart pattern for backend code — causal traces, transactional state, and auto-generated tool descriptions. Everything else is built on this.
The self-explaining stack.
From backend pipelines to AI agents — every run records why it did what it did. Trace any answer back to the exact cause, and prove the fix by replaying it.
Every run records itself as it happens — not reconstructed from logs afterward. One trustworthy trace, and you choose the lens.
1
Collect at traversal time
Reads, writes, and decisions are captured structurally as the run executes — no manual logging.
2
One canonical footprint
The trace is the execution record — built inline, never rebuilt after the fact.
3
You choose the lens
Metrics, narrative, visualization, agent debugging — every tool composes on the same record.
Inline recording is truth. Post-processing is reconstruction.
agentfootprint
agentic frameworkBuild AI agents whose every LLM call traces back to what was injected, who triggered it, when, and how it cached. Context engineering, abstracted. Built on footprintjs.
Explainable UI
for footprintjsVisualize a footprintjs run — the flowchart, the causal trace, and the data at every step. Themeable React components.
Lens
for agentfootprintDebug an agentfootprint run — messages, prompts, tool calls, decision scope, and cost. Built on Explainable UI.
Thinking UI
role replayThe non-developer's view of an agent run — its reasoning and tool loop replayed as an animated, scrubbable story, so anyone can see what it did and where it went wrong. No build step.