Sloth Handover: three lines for standup from an open-weight model

说明与用途

Sloth Handover 将 git 日志转换为三行简洁的站会摘要，使用开源 Gemma 3 27B 模型，无数据持久化，可选本地 Ollama 推理。

问题

Developers spend time summarizing git logs for standup meetings, leading to repetitive and time-consuming effort.

结果

The app transforms a messy git log into three concise standup lines and a minutes‑saved estimate in under 4 s, with no data retention and full privacy control via local inference.

工作流

• Paste or load a git log into the textarea.
• The client sends the log to the /api/handover endpoint.
• The server sanitises the log, caps it at 8,000 characters, and forwards it to the AI provider (OpenRouter or Ollama).
• The model returns a JSON payload with Done, Next, Blocked, and minutes_saved fields.
• The client validates the JSON against a schema and displays the three lines and estimate.
• The app is deployed on Vercel and can run locally with a local Ollama instance.
• The entire process takes ~3.4 s for a 12‑commit sample.
• No data is persisted; the app is stateless.

证据

Sloth Handover turns a messy git log into the three lines you actually say at standup.

来源

No account, no database, no stored history. The log goes to the model, the reply is checked against a schema, and nothing is written down.

来源

The endpoint sanitises the log, caps it at 8,000 characters, rate‑limits eight requests per IP per ten minutes in memory, and answers 400, 429 or 502 with a sentence instead of a stack trace.

来源

I tested the plumbing with a local OpenAI‑compatible stub before spending a real model call: happy path, one‑character log (400), malformed JSON (400), rate limit (429). Then the real call: 12 commits in, three accurate lines out, 3.4 seconds.

来源

The model is the point, not a black box behind a paywall. Gemma 3 27B has open weights, so the same prompt can run through a hosted API for a public demo or through Ollama on a laptop with the network unplugged.

来源