Runlog 2.4: parallel tool calls and step replays

[00] blog

Field notes.

Engineering, research and product notes from the team building Runlog.

fig. 01 / featureddemo

engineeringFeatured

Why every agent needs a verify step

Most agent failures are not model failures. They are missing checks. Here is how we gate every result with evals, and what it cost us in latency.

Mara OkaforSep 24, 2026 · 9 min read

fig. 02
researchMeasuring tool-call reliability at 41M calls a monthSep 18, 2026 · 7 min
fig. 03
productBudgets: stop a run before it costs youSep 12, 2026 · 4 min
fig. 04
customersHow Northwind moved four agents to productionSep 03, 2026 · 6 min

[01] all posts

4 posts

[02] newsletter

Field notes, monthly.

One email. No launches dressed up as insights.