Your agentsworked all night.Watch the cut.
A 60-second reel of what your coding agents did, checked against what they actually ran.
What you get
A reel about a minute long. It opens on the problem you set, walks through the key decisions and the implementation, and ends on the validation that was actually recorded.
Open this reel with its sourcesacme-api-demo, 8 narrated lines
How it reads the night
Paste a repo
Dailies reads last night’s Entire checkpoints: every prompt, command, edit and message.
The agent checks every claim
Each thing an agent said is held against what was actually run. Nothing goes into the reel on its word alone.
You watch, then decide
Watch the reel, ask for changes in your own words, and approve it when it is right.
Every sentence has a source
Put the loupe on any line of the narration and it opens to what it rests on: the transcript entry and the patch. A line with nothing behind it is not narrated as fact.
0:27
The sender now retries 4 times (verified) and waits longer after each attempt.
Transcript entrysession 3 · entry 209 · 02:12
Edit src/webhooks/send.ts +18 −4
Patchsrc/webhooks/send.ts
− const res = await fetch(url, { method: 'POST', body });− if (!res.ok) throw new Error('webhook failed');+ for (let attempt = 0; attempt < 4; attempt++) {+ const res = await fetch(url, { method: 'POST', body, signal: AbortSignal.timeout(5000) });+ if (res.ok) return;+ await sleep(250 * 2 ** attempt);+ }+ throw new WebhookError(url, 4);
Four verdicts
Every claim an agent made leaves the night with one of four pencil marks.
Verified
The record shows it.
Contradicted
The record shows otherwise.
Stale
The code was edited after the last passing run.
Unverified
Nothing in the record either way.
Runs where your agents run
Dailies is a plugin for the Entire CLI.
entire dailies run .