> ## Documentation Index
> Fetch the complete documentation index at: https://docs.fincept.in/llms.txt
> Use this file to discover all available pages before exploring further.

# Credits and limits

> What Fincept MCP requires, what it costs, and the limits that apply to every call.

## Plan

Fincept MCP is included in every Fincept Exclusive plan: **Exclusive**, **Exclusive+** and **Exclusive Pro**. Every tool your plan includes in the terminal is available to your agent; quotas (watchlists, portfolios, history depth) are the same.

## Credits

Most tools are free. Tools that cost credits in the terminal cost the same here, from the same balance:

* The price is in each tool's description, so the agent sees it before calling.
* Charges happen on a delivered result. Failed, empty and refused calls cost nothing.
* Daily spend caps and per-action hourly caps apply as in the terminal (`402 spend_cap_reached`).
* Engine tools are free: Fincept Stats, Forecast, Allocation, Portfolio Optimizer, Quant, Volatility, Panel, TS Forecast, Technicals, Change Points, Regimes, Copula, Multi-Period, Portfolio Lab, Pricing, Scanner and Performance.

The [credits reference](/reference/toolsets/credits) lists every tool that can spend credits. Check your balance with `account_get_credits` and your spending with `account_get_usage`.

## Rate limit

| Limit | Value |
| - | - |
| Requests per account | 120 per minute, bursts up to 40 |
| Request body | 1 MB |

Every MCP message counts as one request, including `tools/list` and discovery calls. Past the limit the server answers HTTP `429` with `Retry-After`; clients retry on their own.

Fincept Scanner's requests to the market data source are also paced across the whole server. When that pace is reached, a scanner tool answers `busy` with `details.retry_after_s`, the seconds to wait; retrying sooner gets the same answer.

## Time limits

| Call | Limit |
| - | - |
| Most tools | 30 seconds |
| Analytics, screeners, portfolio reports, AI answers | 60 to 180 seconds |
| `backtest_run` / `backtest_optimize` | 5 / 10 minutes |
| Fincept Stats | 90 seconds, up to 300 seconds for heavy models |
| Fincept Forecast | 90 seconds, up to 300 seconds for heavy tools |
| Fincept Allocation | 90 seconds, up to 300 seconds for heavy tools |
| Fincept Portfolio Optimizer | 90 seconds, up to 180 seconds for heavy tools |
| Fincept Quant | 90 seconds, up to 300 seconds for heavy tools |
| Fincept Volatility | 90 seconds, up to 300 seconds for heavy tools |
| Fincept Panel | 90 seconds, up to 300 seconds for heavy tools |
| Fincept TS Forecast | 90 seconds, up to 300 seconds for heavy tools |
| Fincept Technicals | 90 seconds |
| Fincept Change Points | 90 seconds, up to 300 seconds for heavy tools |
| Fincept Regimes | 90 seconds, up to 300 seconds for heavy tools |
| Fincept Copula | 90 seconds, up to 300 seconds for heavy tools |
| Fincept Multi-Period | 90 seconds, up to 300 seconds for heavy tools |
| Fincept Portfolio Lab | 90 seconds, up to 300 seconds for heavy tools |
| Fincept Pricing | 90 seconds, up to 300 seconds for heavy tools |
| Fincept Scanner | 75 seconds; `scan_history` and `scan_returns` answer within about 45 seconds and continue in the next call |
| Fincept Performance | 30 to 120 seconds; `perf_rolling` and `perf_bootstrap` 180 seconds, `perf_backtest_overfitting` 300 seconds |

A call past its limit returns an error envelope with code `timeout`.

## Result size

| Item | Value |
| - | - |
| Inline result | About 40 KB; larger results are stored and previewed |
| Stored result | Up to 25 MB, kept 24 hours |

See [results](/guides/results#large-results).


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.