Skip to content
TribalMind

Quick Start

This guide walks you through the core TribalMind workflow: storing knowledge, searching it, and setting up your AI agents to use it automatically.

Prerequisites

Section titled “Prerequisites”
  • TribalMind installed (pip install tribalmind)
  • A project initialized (tribal init)

See Installation if you haven’t done these yet.

Core workflow

Section titled “Core workflow”

  1. Store knowledge

    Use natural language — TribalMind’s LLM parses it into structured memory:

    Terminal window
    tribal remember "redis gives ECONNREFUSED on macOS after sleep — restart the container"

    You can also pipe input:

    Terminal window
    echo "staging DB is on port 5433, not 5432" | tribal remember

  2. Search your memory

    Find relevant knowledge by meaning, not exact keywords:

    Terminal window
    tribal recall "redis connection errors"

    Output shows matched memories ranked by relevance. You can also filter by category:

    Terminal window
    tribal recall --category fix "redis connection errors"

  3. Search across all projects

    Use --all to search every repo in your Backboard account:

    Terminal window
    tribal recall --all "auth token handling"

  4. Browse all memories

    List everything without a search query (no API cost):

    Terminal window
    tribal recall --list

  5. Remove outdated knowledge

    Terminal window
    tribal forget "old redis fix"

    TribalMind will find matching memories and ask for confirmation before deleting.

  6. Check project status

    Terminal window
    tribal status

    Shows your project root, assistant ID, memory count, and LLM provider.

Set up your AI agents

Section titled “Set up your AI agents”

Run tribal setup-agents to generate config files for your AI coding tools:

Terminal window
tribal setup-agents

This writes TribalMind usage instructions into agent config files like CLAUDE.md, .cursorrules, and .github/copilot-instructions.md. Your agents will then automatically recall knowledge when debugging and store knowledge when they discover something useful.

See the Agent Integration guide for details.

JSON output

Section titled “JSON output”

Every command supports --json for machine-readable output — useful for scripting and agent consumption:

Terminal window
tribal recall --json "deployment issues"
tribal status --json

Next steps

Section titled “Next steps”