Everything TeamLaboration does, in detail.
A long-scroll breakdown of every feature in the Solo tier. If you skim only the headlines you'll still get the shape of it.
Every Claude Code session becomes a structured entry.
When a session finishes, TeamLaboration extracts a title, the key topics, decisions made, and the outcome — usually 1–3 sentences. Stored locally in SQLite. Raw logs are never modified or copied.
One narrative across your whole week.
Per-chat summaries are aggregated into a coherent weekly report — themes, accomplishments, and notable decisions. Generated on demand, so the AI cost is bounded. Re-runnable for past weeks.
Ask Claude Code about your own week.
TeamLaboration exposes 13 MCP tools — list weekly reports, search summaries, flag highlights, trigger a re-summarization. Any MCP client can use them. Configuration mutations (API key, polling) are intentionally not exposed.
claude mcp add teamlaboration --transport http --url http://127.0.0.1:7842/mcpSolo tier has no server. Period.
The daemon, the database, and the dashboard all run on your machine. The only outbound call is to Anthropic, using your own API key. No analytics, no telemetry — there is no TeamLaboration server in Solo tier.
Push the report where you actually read.
In-app dashboard at localhost. Markdown or HTML file in a folder you choose. Slack or Teams via webhook. Email via your SMTP credentials. All optional, all configured locally.
Bring the model you trust.
Default is Claude with your API key. The summarizer is abstracted to allow Azure OpenAI, Amazon Bedrock, Ollama, and any OpenAI-compatible endpoint in future versions. Today, Solo ships Claude only.
Sensible defaults. Tunable if you want.
Polls ~/.claude/projects/ every 5 minutes (configurable). Stores 4 weeks of summaries by default; raise or lower the cap as you like. Working-hours / week boundaries default to Mon–Sun. Export to JSON or Markdown before purge.