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

# Limits

> Size, count, and rate limits across Arklex Platform, in one place.

This page collects the limits that apply across Arklex Platform. Each section links to the page that explains the feature. On cloud, token allowances depend on your plan. See [Billing](https://docs-platform.arklex.ai/settings.md#billing). On self-hosted deployments, your operator can change the API and tracing rate limits.

***

## Scenarios

| Limit | Value |
| - | - |
| Scenarios written per assistant run | 100 |
| Conversation history upload | 256 MB, `.json` or `.jsonl` |
| Scenario import file | 10 MB, `.json` or `.csv` |
| Scenario name | 100 characters |
| Goal | 1,000 characters |
| User profile | 3,000 characters |
| Agent context | 2,000 characters when you import or write scenarios, 3,000 when you edit one or use the API |
| Knowledge passages per scenario | 5, each with a name up to 120 characters and content up to 2,000 |
| User goal checklist items | 1 to 4, at least one Required. Each description up to 500 characters. |
| Assertions per scenario | 50, including at most one **Tool calls** assertion |

See [Scenarios](https://docs-platform.arklex.ai/scenarios.md).

***

## Simulations

| Limit | Value |
| - | - |
| Conversations per scenario | 10 |
| Max turns per conversation | 10 |
| Max concurrent conversations | 50 (default 10) |
| Conversations per simulation | 2,000 |
| Trace wait time | 10 to 300 seconds (default 60) |
| Simulations pending or running at once | 2 on the cloud Free plan. No limit on other plans or on self-hosted. |

See [Simulations](https://docs-platform.arklex.ai/simulations.md).

***

## Knowledge

| Limit | Value |
| - | - |
| Document upload | 10 MB per file |
| Links crawled per domain | 1 to 1,000 (default 50) |
| Folder depth | 6 levels |

See [Knowledge](https://docs-platform.arklex.ai/knowledge.md).

***

## Metrics and MCP servers

| Limit | Value |
| - | - |
| Custom metric name | 100 characters |
| Custom metric description | 512 characters |
| MCP tool calls per metric | Turn scope: 5 per turn and 20 per conversation. Conversation scope: 5 in total. |
| MCP connection test timeout | 10 seconds |
| MCP tool call timeout | 30 seconds |
| Tools listed per MCP server | 500 |

See [Metrics](https://docs-platform.arklex.ai/metrics.md) and [MCP servers](https://docs-platform.arklex.ai/mcp-servers.md).

***

## Annotations and settings

| Limit | Value |
| - | - |
| Comment length | 2,000 characters |
| Profile photo | 2 MB, JPEG, PNG, or WebP |
| API key name | 100 characters |
| Invitation lifetime | 7 days |

See [Annotations](https://docs-platform.arklex.ai/annotations.md) and [Settings](https://docs-platform.arklex.ai/settings.md).

***

## API and tool call tracing

| Limit | Value |
| - | - |
| API requests per key | 500 per minute, with bursts up to 600 |
| Connection tests per key | 10 per minute |
| Idempotency key lifetime | 24 hours after the first attempt completes |
| Trace requests per organization | 600 per minute |
| Trace data per organization | 64 MiB per minute |
| Trace request size | 4 MiB compressed. Uncompressed, 16 MiB for protobuf and 4 MiB for JSON. |
| Spans per trace request | 2,000 |
| Trace data kept per turn | 2,000 spans and 16 MiB |
| Tool call arguments, result, and error | 64 KiB, 64 KiB, and 8 KiB. Longer values are cut. |

See [API overview](https://docs-platform.arklex.ai/api.md) and [Tool call tracing](https://docs-platform.arklex.ai/tool-tracing.md).


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