TEXT TO QUANT

Backtest a trading idea written in plain English from an AI agent.

Opens Agent Studio, where connecting is one click. The connector URL below works in any MCP client.

Paste it into any MCP client. Setup docs

What the TEXT TO QUANT connector does

The TextToQuant MCP server lets Claude turn a trading idea written in plain English into a backtest on real crypto, forex and equity data, then grade it, stress-test it and export it to Pine Script. It signs in to your TextToQuant account, and each run spends your plan's backtest credits. It never places a trade or touches a brokerage account.

Verified connector

Listed by Anthropic as a partner connector in its Connectors Directory.

Connection checked by Agentman on .

Anthropic states this reflects the level of review a connector received, not a security audit.

TEXT TO QUANT tools (78)

  • analyze_conditions
  • cancel_backtest
  • cancel_portfolio_sweep
  • compare_backtests
  • create_scan_alert
  • custom_benchmark
  • delete_indicator
  • delete_scan_alert
  • describe_capabilities
  • diff_iterations
  • edit_backtest
  • explain_grade
  • explain_parse
  • export_pine
  • filter_by_context
  • get_analysis
  • get_backtest
  • get_backtest_data
  • get_candles
  • get_context_trades
  • get_indicator
  • get_indicator_preview
  • get_iterations
  • get_notification_channels
  • get_overfit_verdict
  • get_plans
  • get_portfolio
  • get_portfolio_analysis
  • get_portfolio_sweep
  • get_profiler
  • get_regime_edge
  • get_usage
  • lint_indicator
  • list_backtests
  • list_context_dimensions
  • list_examples
  • list_indicators
  • list_markets
  • list_portfolios
  • list_scan_alerts
  • manage_webhooks
  • market_regime
  • market_sectors
  • organize_history
  • parse_portfolio
  • parse_strategy
  • preview_indicator
  • preview_js_indicator
  • preview_py_indicator
  • publish_to_gallery
  • refine_context
  • rephrase_strategy
  • rsps_matrix
  • run_backtest
  • run_portfolio
  • run_portfolio_sweep
  • run_profiler
  • save_csv_indicator
  • save_js_indicator
  • save_pine_indicator
  • save_py_indicator
  • saved_scans
  • scan_market
  • share_backtest
  • share_portfolio
  • show_analysis
  • show_backtest
  • show_context
  • show_portfolio
  • start_analysis
  • token_multitf
  • token_stats
  • update_scan_alert
  • validate_indicator
  • validate_js_indicator
  • validate_py_indicator
  • workspace_summary

Tool names from Anthropic's directory listing. This server requires sign-in, so we could not read tool descriptions or parameter schemas.

Limits

  • Credits are the cost. Runs, sweeps, grids and portfolios bill backtest credits. A 5×5 parameter grid costs 25, per TextToQuant's credit table. Reading, parsing, sharing and exporting are free.
  • Plan prices, from TextToQuant's pricing summary. Free, with five backtests for the life of the account; Pro $47 a month; Power $99; Quant $499; Enterprise on request. Annual billing is lower.
  • Not advice, not trading. Output is historical simulation, which TextToQuant says is not a recommendation. The connector places no orders and connects to no broker.
  • Plan-gated tools fail at call time. TextToQuant says tools a plan lacks return a plan_required error instead of being hidden.
  • Rate limits. TextToQuant documents 240 requests a minute per IP plus a per-account ceiling set by plan, and a daily budget of candle bars.
  • Consent does not itemise. The OAuth grant asks only for identity scopes. TextToQuant says it enforces read, run and write scopes per tool on its side, which the consent screen does not show.
  • Spend confirmation is conditional. The confirm-before-spend prompt applies "when enabled on the server" and only in clients that support elicitation.
  • We could not read the tool list. The server returned HTTP 401 before tools/list, so names come from the directory and behaviour from TextToQuant's documentation. We called no tool.

Sources

Use it in an agent

Put TEXT TO QUANT to work.

Connect TEXT TO QUANT once and your agents call these tools on their own: on a schedule, in a workflow, with nobody at the keyboard.

Call (650) 285-1019, Agentman.Our AI receptionist answers.