TablaFocusMCP

TablaFocusMCP is a MCP server focused on tabla learning workflows. It unifies glossary lookup, composition design and validation, certification preparation, practice planning, and taal explanation into one consistent interface for AI assistants and serious learners. Alongside tools, it also exposes MCP resources (readable datasets) and prompts (guided workflows).

How To Use

Once tablafocus is installed in your MCP client, chat naturally about what you want to practice or build. The assistant can explain taals, generate compositions, create quizzes/mocks, and plan practice weeks.

Try prompts like these:

1. Explain teental for a beginner, including vibhag structure, sam, and khali.
2. Generate a valid 1-cycle tihai in teental (chatusra) and show the beat-by-beat mapping.
3. Create a weekly tabla practice plan for 45 minutes/day, 5 days/week, focused on clarity and layakari.
4. Build a 15-question ABGMVM Madhyama Pratham mock test with answer key and short rationales.
5. I have a tihai idea for teental. Help me refine it so it resolves cleanly to sam in 1 cycle.
6. Transpose this teental tihai into rupak while preserving the structural feel as closely as possible.

Workflow prompt templates are also available:

Prompt name	What it guides	Inputs
cert_prep_plan	Certification workflow (catalog -> mock -> plan)	board, certification_level, days_per_week, minutes_per_day
weekly_practice_reset	Weekly reset workflow after missed sessions or fatigue	goals (semicolon-delimited), daily_minutes, days_per_week, optional missed_days, completed_minutes, fatigue
exam_week_plan	Focused 7-day exam prep workflow	board, certification_level, daily_minutes, optional weak_areas, fatigue
missed_week_recovery	Recovery workflow after a disrupted practice week	goals (semicolon-delimited), daily_minutes, days_per_week, optional missed_days, completed_minutes, fatigue
composition_polish	Iterative composition draft -> validate -> refine flow	taal, form, jati, optional cycles, optional polish_rounds

Core Tools

All tools return a common envelope with meta and data.

Tool	What it does	Required inputs	Optional inputs	Output highlights
glossary_lookup	Finds glossary terms and definitions	None	term, category, limit (1-100)	Matching entries, categories list, total matches
compose_builder	Builds mathematically valid compositions	taal, form, jati (tisra|chatusra|khanda|misra)	cycles (1-12)	Composition equation, parameters, timeline segments, alternatives
composition_transposer	Transposes a valid composition into a new taal/jati context	source (taal, form, jati, cycles, composition_input), target (taal, jati)	target.cycles, preserve_mode=shape_ratio	Chosen transposed composition, scale factor, preservation report, alternatives, warnings
certification_catalog	Lists certification tracks and level breakdowns	None	board, certification_level	Board/level catalog with papers, categories, objectives, references
assessment_builder	Creates quizzes and certification mocks	mode (practice_quiz|cert_mock)	count (1-100, default 10), seed, board, certification_level, taal	Questions, answer key with rationale, rubric, optional certification reference
practice_coach	Generates adaptive weekly practice plans	goals (array), availability (object)	profile_id, availability.daily_minutes (1-600), availability.weekly_minutes (1-4000), availability.days_per_week (1-7), week_context.missed_days (0-7), week_context.completed_minutes (0-4000), week_context.fatigue (low|medium|high)	Weekly target, daily targets, per-day sessions, adaptive adjustments
taal_catalog	Returns full catalog or one taal detail	None	taal_id	Taal structure, vibhag, sam/khali, clap-wave, counting guidance, theka
composition_validator	Validates composition equation and timeline checks	taal, form, jati, cycles (1-12), composition_input	composition_input.P, G, M, g, optional detailed segments	is_valid, failure reasons, equation/timeline/segment checks
explain_taal	Compatibility alias for taal explanation	taal	None	Explanation payload sourced from canonical taal_catalog data

How To Install

Run directly from npm:

npx -y tablafocus-mcp@latest

Codex CLI:

codex mcp add tablafocus -- npx -y tablafocus-mcp@latest

Claude Code:

claude mcp add -s user tablafocus -- npx -y tablafocus-mcp@latest

JSON-based clients (Claude Desktop, Cline, VS Code Copilot):

{
  "mcpServers": {
    "tablafocus": {
      "command": "npx",
      "args": ["-y", "tablafocus-mcp@latest"]
    }
  }
}

Cursor:

{
  "name": "tablafocus",
  "command": "npx",
  "args": ["-y", "tablafocus-mcp@latest"]
}

Documentation

• Client setup
• MCP reference
• Changelog
• Contributing guide
• MIT License