Session Management

Sessions are persisted to SQLite at ~/.yeaboi/data/sessions.db. Every terminal session gets a unique ID (new-<8hex>-<YYYY-MM-DD>) and a human-readable display name derived from the project slug (todoapp-2026-03-19). Team analysis profiles are stored in the same database.

Directory structure

~/.yeaboi/
  data/
    sessions.db         # SQLite — planning sessions, analysis profiles, token usage
  exports/
    analysis/           # HTML + Markdown analysis reports
    planning/           # HTML + Markdown planning exports
  logs/                 # All logs rotate at 2 MB with 3 backups, one shared format
    tui/                # Main application log (yeaboi.log) — always on
    analysis/           # Analysis mode log + per-analysis text reports
    planning/           # Per-planning-session logs ({session-id}.log)
    standup/            # Daily Standup log (standup.log, incl. scheduled runs)
    retro/              # Retro mode log (retro.log)
    performance/        # Performance mode log (performance.log)
    reporting/          # Reporting mode log (reporting.log)
  .env                  # API keys and credentials

Resume a session

yeaboi --resume            # interactive picker
yeaboi --resume latest     # most recent session
yeaboi --resume <id>       # specific session ID

Resumed sessions pick up exactly where you left off — mid-questionnaire, mid-review, or between pipeline stages.

List sessions

yeaboi --list-sessions

Shows a table with project name, date, last completed step, and session ID.

Delete sessions

yeaboi --clear-sessions

Interactive picker to delete one session or clear all.

Auto-pruning

Sessions older than 30 days are auto-pruned at startup. Configure via SESSION_PRUNE_DAYS in .env (set to 0 to disable).

Usage Page

The Usage page tracks token consumption across all sessions, accessible from the main menu.

Metric	Description
Session tokens	Input and output tokens for the current session
Lifetime tokens	Cumulative total across all sessions, persisted in SQLite
Per-provider breakdown	Tokens split by LLM provider (Anthropic, OpenAI, Google, Bedrock)
Session history	Recent sessions with their token counts and timestamps
Cost estimate	Approximate cost based on provider pricing

Token usage is tracked automatically via track_usage() on every LLM call and persisted to the token_usage table in SQLite. The page uses a dedicated amber colour theme to distinguish it from Planning (blue) and Analysis (green).

Settings Page

The Settings page provides a read-only view of your current configuration and a shortcut to the setup wizard.

Section	What it shows
LLM Provider	Active provider and model (e.g., Anthropic / claude-sonnet-4)
API Keys	Configured keys with values masked (e.g., `sk-ant-*...*abc`)
Issue Tracking	Jira and/or Azure DevOps connection status, org URL, project
Version Control	GitHub token status
Storage	The data directory (`YEABOI_HOME`) everything is written under

From the Settings page you can launch the setup wizard to reconfigure providers, API keys, and integrations, cycle the log level, and change the data directory (with an offer to move your existing sessions, exports, and logs to the new location). The page uses a grey colour theme.