Builder Demo Artifact
workledger turns a run log into a work log.
This static artifact is based on the bundled agent-workflow demo. It shows the exact moment the project becomes intuitive: 14 raw spans turn into 3 work units, covering maintenance, product development, and internal automation, with one item queued for review instead of being forced into certainty.
Before
What the trace layer gives you
- LLM calls, retrievers, tools, guardrails, and human review spans
- Token counts, latency, and direct cost by span
- No compact answer to “what work happened?”
After
What workledger gives you
- Reviewable work units with titles and project context
- Cost by work category, evidence, and review state
- A review queue for the one item that still needs judgment
Work Units
Three outputs that a builder can reason about
| Title | Category | Cost |
|---|---|---|
| Implement orchestration dashboard for customers | external_product_development | $0.0828 |
| Patch customer API timeout regression | maintenance_bugfix | $0.0483 |
| Automate release checklist workflow | internal_use_software | $0.0471 |
Review Queue
The ambiguous item stays reviewable
Automate release checklist workflow
internal_use_software -> pending review
Confidence: 0.768
Source spans: 3
Compression: 1.50x
Maintenance Contrast
Patch customer API timeout regression
This is the kind of work observability tools can describe technically but not semantically. workledger rolls six spans into one maintenance work unit with clear cost, evidence, and trace lineage.
maintenance workDevelopment Contrast
Implement orchestration dashboard for customers
Five spans become one external product development unit with a higher blended cost. The useful question becomes review and confidence, not just trace inspection.
external product developmentCompression Story
What builders should remember
The best proof point in this demo is not a dashboard. It is the reduction from 14 spans to 3 reviewable units of work. That smaller unit is what makes cost attribution, evidence review, and downstream business analysis feel obvious instead of aspirational.