PAPAYA Backtests

PAPAYA Backtests turns finalized historical research summaries into consistent, searchable reports in PAPAYA's shared research library.

Read the current report format, check the connected PAPAYA account, validate an approved compact JSON package, and publish it with a link back to the report. An approved PAPAYA account with contributor access is required to validate or publish.

MCP connector

Server URL: https://api.papayaops.com/mcp
Login: API key
Protocol: MCP 2025-11-25
Instructions: Use PAPAYA tools for its current backtest guide and report schema. Refine research with the user. After every update ask whether these finalized results are ready to submit to PAPAYA. Only explicit approval for the current version permits transmitting research, including validation. Connection approval alone is not report approval. Summarize locally; the entire request must be under 2 MiB. Authenticate using the client's PAPAYA sign-in flow; never ask for passwords or API keys in chat. PAPAYA accepts standards 1.0, 1.1 and 1.2. New publications require numeric Profit Factor, Win Rate and Sharpe, a tested universe, test dates and a plain-language strategy introduction. Undefined statistics use 0 with an explicit reason in 1.2. Scatter plots are supported; 1.2 adds box plots and typed evidence metadata. At the start of a submission workflow call and follow its current compatibility_guidance; read the live schema/guide when the contract fingerprint is unknown or changed. Resolve tool argument mismatches automatically first: read the live schema, inspect the actual submission envelope and use a host reload/rediscovery capability once only if it is actually exposed. PAPAYA cannot itself refresh ChatGPT's imported tools. Ask for user intervention only if recovery remains blocked. Keep the exact approved package and idempotency key; a transport-only retry needs no renewed publication approval. Never downgrade, discard evidence or claim an unperformed refresh.

Tools

Reads data4

Changes data1

Skills

None in the store listings or on the server

MCP connector

Server URL

https://api.papayaops.com/mcp

API key

Protocol

MCP 2025-11-25

Instructions

Use PAPAYA tools for its current backtest guide and report schema. Refine research with the user. After every update ask whether these finalized results are ready to submit to PAPAYA. Only explicit approval for the current version permits transmitting research, including validation. Connection approval alone is not report approval. Summarize locally; the entire request must be under 2 MiB. Authenticate using the client's PAPAYA sign-in flow; never ask for passwords or API keys in chat. PAPAYA accepts standards 1.0, 1.1 and 1.2. New publications require numeric Profit Factor, Win Rate and Sharpe, a tested universe, test dates and a plain-language strategy introduction. Undefined statistics use 0 with an explicit reason in 1.2. Scatter plots are supported; 1.2 adds box plots and typed evidence metadata. At the start of a submission workflow call and follow its current compatibility_guidance; read the live schema/guide when the contract fingerprint is unknown or changed. Resolve tool argument mismatches automatically first: read the live schema, inspect the actual submission envelope and use a host reload/rediscovery capability once only if it is actually exposed. PAPAYA cannot itself refresh ChatGPT's imported tools. Ask for user intervention only if recovery remains blocked. Keep the exact approved package and idempotency key; a transport-only retry needs no renewed publication approval. Never downgrade, discard evidence or claim an unperformed refresh.