Troubleshooting & FAQ
This page answers common questions about Generate Enterprise and gives fixes for the most frequent issues — supported formats, models, data handling, and login/chat/upload/connector problems.
Which AI models does Generate Enterprise support?
Section titled “Which AI models does Generate Enterprise support?”Iterate Generate is designed to give enterprises secure, governed Private AI. It starts with a powerful Private AI foundation, enabling organizations to run LLMs securely (Iterate finetuned or Iterate Recommended Open Weight Models) within their own infrastructure—on-premises, in their data center, in a VPC, at the edge, or across sovereign environments.
At the same time, Generate gives enterprises an option to connect to the broader LLM ecosystem when needed. It integrates with frontier and commercial models through API-based and service profiles, including OpenAI, Anthropic, Mistral, Cohere, and other providers, as well as cloud model platforms such as AWS Bedrock. Generate can also connect through OpenRouter and other gateways/exchanges, giving organizations access to a broad and continuously evolving selection of models without redesigning their application and agents.
Your administrators decide which models are used, so users do not select a model per chat.
What file formats and size limits does the Knowledge Base support?
Section titled “What file formats and size limits does the Knowledge Base support?”Generate accepts PDF, Word (DOC/DOCX), PowerPoint (PPT/PPTX), Excel (XLS/XLSX/XLSB), CSV, plain text (TXT), Markdown (MD), and images (PNG, JPG/JPEG, WEBP, SVG). Each file can be up to 200 MB, with up to 50 files per upload. Standalone images are supported, and images, tables, and charts inside documents are read in Precision mode.
How is organization data handled?
Section titled “How is organization data handled?”Generate is self-hosted in the organization’s own environment, so documents and data stay under the organization’s control. Data is isolated per workspace, sensitive fields are encrypted at rest, and administrators control access (by role and team) and data-retention policies.
Why is the AI giving incorrect or irrelevant answers?
Section titled “Why is the AI giving incorrect or irrelevant answers?”Answer quality depends on the Knowledge Base and the question. Confirm the relevant documents are uploaded and finished processing (status Completed), remove outdated documents, be specific, and try rephrasing. If using a project, confirm the document is in the correct project. For citations, use Precision processing mode (Express mode returns answers without citations).
How do you fix login and access problems?
Section titled “How do you fix login and access problems?”Verify the email and password, use “Forgot password” to reset, and check with an administrator in case the account was not yet activated. If a feature is missing, an administrator may not have enabled it for your workspace or role.
How do you fix document upload and processing issues?
Section titled “How do you fix document upload and processing issues?”Confirm the file format is supported and under 200 MB (max 50 files per upload). If a document stays in Processing, wait for large files, then cancel and re-upload or retry the extraction. If the AI does not use a document, confirm its status is Completed and that it is in the correct project.
How do you fix connector issues?
Section titled “How do you fix connector issues?”Sign in with an account that has the right permissions, confirm an administrator has enabled the connector for your workspace, and try disconnecting and reconnecting. After connecting, allow a few minutes for data to sync.
Sources and references
Section titled “Sources and references”- Full customer documentation:
support/faqandsupport/troubleshootingon the Generate Enterprise docs site.