GradientAI Platform Limits

Platform Limits

• You cannot use doctl, the official DigitalOcean CLI, to manage GradientAI Platform resources.

• To manage compute resources and ensure fair resource distribution, GradientAI Platform has limits on resource creation and model usage:

  • Teams have a daily limit on the number of agents they can create.

  • Teams have limited number of tokens available for agents to use. We allocate a predetermined amount of tokens for each model your team uses, and each agent on your team draws tokens from that model’s amount of tokens.

  • Teams have limited number of tokens available for each model tested in the Model Playground. These tokens also replenish every 24 hours.

  Tokens for agents and tokens for the Model Playground both replenish every 24 hours. For example, tokens used at 9:05 on Wednesday replenish at 9:05 on Thursday.

• We set dynamic resource limits to protect our platform against bad actors, so you cannot view your resource or token limits. You can request an increase by contacting support.

Foundation Model Limits

• You cannot bring your own models to use for AI agents or serverless inference. You can see the models we offer on our model overview page.

Serverless Inference Limits

• Serverless inference supports the two to three most recent stable versions of each model to ensure consistent performance and reliable maintenance. For the list of supported models and versions, see the available model offerings.

• Serverless inference model endpoints support OpenAI-compatible request formats but may not be compatible with all OpenAI tools and plugins.

• Serverless inference provides access to commercial models, but not all model-specific features are supported. For example, features like Anthropic’s extended thinking are not available.

• OpenAI models accessed through serverless inference not support zero data retention. If your use case requires strict data privacy or compliance, consider using a different model or contact support for guidance.

• You cannot access the OpenAI o1 model through serverless inference. All other supported models are available as model endpoints.

Agent Limits

• You cannot view the sources used to generate a response in the chatbot interface.

• You cannot access agent tracing data through the API. Tracing is only available through the Agent Playground and an agent’s Observability page in the control panel.

• You cannot define custom fields or metadata for agent tracing. Traces only include information such as inputs, outputs, token usage, processing time, and resource access.

• Agent tracing does not display routing data for other agents accessed during a request.

Knowledge Base Limits

• You cannot edit attributes of a knowledge base using the DigitalOcean API. Instead, edit your knowledge base using the DigitalOcean Control Panel.

• You cannot change embedding models after creating a knowledge base.

• For web crawling data sources, the crawler indexes up to 5500 pages and skips inaccessible or disallowed links to prevent excessively large indexing jobs.

• You cannot currently re-index a previously crawled seed URL. To re-index the content, delete the seed URL, and then add it again to start a new crawl.

• You cannot change the OpenSearch database for an existing knowledge base. To use a different database, create a new knowledge base with the same data sources and choose the new OpenSearch database during setup.

• You cannot re-index specific data sources within a knowledge base. To re-index any changed data sources, you need to re-index all the data sources.

Guardrails Limits

• You cannot customize detection rules for guardrails.

Functions Limits

• We only support web functions for function routing from agents.

• If you have a public agent that calls a private function, anyone with the function’s URL can call the private function. We recommend setting your function to Secure Web Function to enable authentication.

Agent Evaluations Limits

• Each test case dataset can contain up to 500 prompts. If the dataset includes more than 500 prompts, only the first 500 are used in the evaluation.

• Evaluation runs have a 10,000 token cap across all prompts and responses combined.

• You cannot average scores across multiple metrics, so each test case must have one star metric to determine overall performance.

• You cannot configure thresholds for non-star metrics.

• You cannot adjust the number of judges used in your agent evaluations.

• You cannot delete test cases. Instead, you can archive the test cases to hide them from your list of test cases. You can also delete your workspace, which deletes all associated test cases and runs permanently.

• We do not support synthetic dataset generation. You must upload your own dataset.

• You cannot edit datasets through GradientAI Platform. Instead, edit the dataset and then reupload it.