Google Gemini Document Upload Limits: 10-File Caps, Sizes, and Workarounds

Google Gemini restricts web interface uploads to 10 files per prompt at 100 MB per document, as documented in Google's official Gemini Apps support specifications on October 10, 2026. This dual boundary means that whether you submit brief text notes, research PDFs, or complex spreadsheets, the web interface rejects any individual document larger than 100 MB or any prompt containing an eleventh file.

The Google Gemini document upload limit restricts web interface users to a maximum of 10 attached files per prompt up to 100MB each, while the Google Gemini File API allows individual files up to 2GB with a 20GB per-project storage cap and 48-hour retention. Knowing where these boundaries lie prevents unexpected job failures when analyzing technical manuals, financial statements, or legal records.

Google enforces different upload constraints depending on the access tier, model deployment, and protocol:

• Gemini Web Interface (Free Tier): Accepts up to 10 files in a single prompt. Documents, spreadsheets, and text files are capped at 100 MB per file. Video files can reach up to 2 GB with a maximum duration of 5 minutes. Audio files are limited to 10 minutes.
• Gemini Advanced (Google AI Pro and Google AI Ultra Plans): Retains the standard 10-file attachment limit and 100 MB document size cap. The paid subscription increases video duration allowances to 1 hour and audio analysis to 3 hours.
• Google Workspace Accounts (Work and School): File uploads follow consumer size limits but are governed by organizational access policies. Connecting to Google Drive requires the Workspace administrator to enable Gemini Apps access.
• Gemini Files API (Developer Tooling): Intended for programmatic multimodal ingestion in Google AI Studio and Vertex AI. Supports files up to 2 GB each with a project storage limit of 20 GB. Uploaded files persist for 48 hours before automatic deletion.
• Inline API Payloads: Passing file bytes directly as base64-encoded strings inside prompt requests imposes a 100 MB total payload ceiling per API call.

The following comparison table outlines the documented limits across Gemini environments as of October 2026:

Specific constraints apply when uploading compressed archives or project directories. Uploading a folder through the web interface unpacks its contents, with every individual file counting toward the 10-file maximum. If a folder holds 15 small text files, the upload fails. Standard ZIP archives are capped at 100 MB, cannot contain audio or video files, and cannot hold more than 10 internal items. Dedicated code folder uploads and connected GitHub repositories allow up to 5,000 files with a combined maximum archive size of 100 MB.

• Gemini Gems File Limit: 10-File Upload Cap and Knowledge Base SolutionsThe Gemini Gems file limit caps custom knowledge bases at 10 reference files and 100 MB per document. While attached...
• CSV File Size Limits: AI Assistant Upload Caps, Parsing Ceilings, and FixesCSV file size limits vary from strict row boundaries in spreadsheets to memory and token constraints in AI chat...
• OpenAI Assistant File Limits: Upload Caps, Size Limits, and SolutionsOpenAI assistant file limits balance responsiveness against overhead, restricting direct attachments while capping...
• Perplexity Spaces File Limit: Upload Caps, Supported Formats, and Storage WorkaroundsPerplexity Spaces enforces a 25MB per-file upload limit and a 50-document ceiling per project on Pro accounts. While...
• NotebookLM PDF Limit: Word Caps, Page Limits, and Gemini Notebook RulesGoogle's Gemini Notebook, formerly NotebookLM, enforces a strict limit of 500,000 words and 200MB per uploaded PDF,...
• Google Gemini PDF Limits: File Sizes, Page Caps, and Parsing FixesThe Gemini PDF limit enforces distinct document ceilings across environments: 100MB and 10 files per prompt in web...

More on this subject: Agent File and Document Workflows (298 guides)

Clearing the 100 MB file size threshold does not guarantee that Gemini can process your document. File size measures disk storage in megabytes, whereas model comprehension is constrained by context window token capacity. A dense 40 MB text export can contain several million words, exceeding the processing capacity of any frontier model. Conversely, an 80 MB PDF containing scanned diagrams might contain only a few thousand words but will consume visual tokens for each rendered page.

Gemini models feature industry-leading context windows, but those windows remain absolute boundaries:

• Gemini 1.5 Pro: Supports an input context window of 2,097,152 tokens.
• Gemini 1.5 Flash: Provides an input context window of 1,048,576 tokens.
• Gemini 2.0 Flash: Maintains a 1,048,576 token input context window.

When a user uploads a PDF document, Gemini parses the content through two distinct tokenization pipelines:

• Digital Text Layers: Machine-readable text embedded in PDFs is extracted directly. In standard English prose, one token represents roughly four characters, or approximately 0.75 words. A 500-page plain text document contains roughly 200,000 tokens, which fits comfortably within Gemini's input window.
• Scanned Pages and Visual Elements: Scanned documents, technical drawings, charts, and embedded photographs are converted into visual tiles. Gemini standardizes image processing by dividing pages into 384 by 384 pixel patches, each consuming 258 tokens. A 300-page scanned legal filing consumes 77,400 visual tokens before accounting for optical character recognition of the text content.

The Gemini File API provides developer infrastructure for large multimodal files, but it is engineered as an ephemeral staging cache rather than a persistent document repository. Google's official documentation defines the exact storage boundaries: "The Files API lets you store up to 20 GB of files per project, with a per-file maximum size of 2 GB. Files are stored for 48 hours."

Google offers the Files API at no cost across all regions where the Gemini API operates. Because storage is unmetered, Google enforces automatic deletion after 48 hours to prevent the service from becoming an unmanaged dumping ground for terabytes of media. Developers can query file metadata or manually delete files before the 48-hour expiration using the files.delete method. However, user-uploaded files cannot be downloaded back through the API. The files.download method is restricted to model-generated assets, such as rendered video files.

This 48-hour retention limit introduces serious obstacles for enterprise production systems:

• Broken Agent Loops: Autonomous agents and automated data pipelines that store file resource names in relational databases fail when the 48-hour window expires, returning HTTP 404 errors on subsequent runs.
• Bandwidth Overhead: Ingesting a 15 GB corporate documentation library requires re-uploading the entire dataset every two days. For distributed teams, this generates continuous network overhead and introduces latency into scheduled analysis jobs.
• The Project Quota Ceiling: When an organization needs to analyze 35 GB of historic contracts or medical scans, the Files API rejects additional uploads once the 20 GB project cap is reached. Engineers must write custom garbage collection scripts to prune older files.
• Siloed Developer Keys: Files uploaded to the Files API are bound to a single Google Cloud project and API key. Non-technical team members, auditors, and external collaborators cannot inspect, organize, or review the underlying documents.

Analyze Large Document Repositories Without Gemini Upload Caps

Connect your AI assistants to persistent Fastio workspaces with automated hybrid search and consolidated MCP tools. Monthly plans start with a 14-day free trial, credit card required.

Start 14-Day Trial

Violating Gemini upload boundaries produces specific interface warnings and API errors. Identifying the root cause allows you to apply the correct technical fix rather than guessing at file formatting issues.

Common upload errors include:

• "Delete data to upload file": This error occurs in the Gemini web application when your internal Gemini Apps Activity storage is full. This storage allocation is independent from your Google Drive or Google One storage. Even if your Google One account has 2 TB of free space, a full Gemini Activity quota blocks further document uploads. To resolve this in personal accounts, open Gemini Apps Activity and delete historical conversations that contain heavy attachments. In Workspace accounts, your administrator must clear activity history or adjust retention policies.
• File size ceiling warning: Displayed in the Google Gemini browser interface when attempting to upload a document, spreadsheet, or archive that exceeds 100 MB. To bypass this, split multi-chapter PDFs into smaller files or convert dense spreadsheets to compressed CSV formats.
• "Too many files in prompt": Appears when dragging more than 10 files into the chat box or uploading a folder containing more than 10 files. You must combine text into single files or upload documents sequentially in separate prompts.
• "Upload limit reached" (Rolling Usage Limits): Gemini applies rate limits to document analysis over rolling windows. If you submit several maximum-sized PDFs within minutes, the web interface temporarily disables attachments. Waiting one to two hours resets the rolling window.
• "400 Invalid Argument: File size exceeds 2 GB": Returned by the Gemini Files API when a file exceeds 2 GB or the project crosses the 20 GB threshold. Calling files.list and deleting expired files resolves the issue.
• Context Window Truncation: If a document's total tokens exceed the model's context capacity, Gemini either displays a processing error or truncates earlier text, leading to incomplete answers.

Organizations managing complex document libraries, such as compliance repositories, product technical manuals, or legal litigation archives, cannot rely on 10-file chat limits or 48-hour temporary API storage. Five architectural patterns allow teams to analyze large corpora without running into upload ceilings:

1. Local File Splitting and Extraction: Developers use local scripts with libraries like pypdf or pdfcpu to split 1,000-page files into smaller segments. While simple, this approach strips visual layouts, loses page context, and requires local computational resources.
2. Google Drive Integration: Gemini web users can connect Google Drive to query stored files. However, Workspace accounts require administrative activation, attachments remain subject to per-prompt token limits, and multi-file cross-referencing remains constrained.
3. Pre-Signed Object Storage URLs: Developers store documents in Google Cloud Storage or Amazon S3 buckets and pass temporary pre-signed URLs to the Gemini API. While bypassing the 2 GB Files API limit, raw object storage provides no search indexing, forcing the model to ingest complete files on every request.
4. Custom Vector Databases: Engineering teams construct Retrieval Augmented Generation (RAG) pipelines using vector databases like Pinecone or Qdrant. While functional, building and maintaining custom chunking logic, embedding generation, vector indexes, and sync routines requires dedicated developer maintenance.
5. Intelligent Cloud Workspaces via Model Context Protocol: A cleaner approach connects Gemini directly to a persistent, intelligent cloud workspace. Instead of attaching files to prompts or re-uploading documents every 48 hours, teams store their complete document library in a shared workspace on Fastio workspace storage for agents.

Fastio delivers an intelligent workspace platform designed for human-agent collaboration. Files uploaded to a workspace are automatically indexed for hybrid search, which combines full-text keyword indexing, semantic understanding, and search by metadata value. Rather than forcing an LLM to read 500 pages of text, an assistant queries the workspace and retrieves only the relevant paragraphs with citations.

Workspaces provide shared org-owned repositories, per-file version history, and a detailed activity log that records modifications across team members and agents. Documents can be imported from Google Drive, Dropbox, Box, and OneDrive. Fastio supports Cloud Sync for Google Drive, Dropbox, Box, and OneDrive on a schedule or on demand.

Connecting Gemini CLI or custom agents to Fastio takes minutes using the remote Model Context Protocol (MCP) server over Streamable HTTP. In the Gemini CLI configuration file (settings.json), add the server under mcpServers:

{
  "mcpServers": {
    "fastio": {
      "httpUrl": "https://mcp.fast.io/mcp/code"
    }
  }
}

Sign in with OAuth in your browser to complete authentication. When an agent needs information from an indexed 500 MB manual, it calls the search tool. Fastio returns the exact passages required for inference, consuming only 1,500 prompt tokens instead of exhausting context limits. Review setup steps in the Fastio MCP setup documentation.

For structured document processing across contracts, receipts, or medical forms, teams deploy Metadata Views. Metadata Views extracts typed fields (Text, Integer, Decimal, Boolean, URL, JSON, Date & Time) from unstructured documents into an interactive database without manual template training.

Monthly plans start with a 14-day free trial, which requires a credit card. Plans are Starter at $9.99/mo (3 seats, 250 GB, 5 workspaces, 100,000 credits a month), Business at $49.99/mo (10 seats, 5 TB, 50 workspaces, 600,000 credits a month), and Enterprise at $199.99/mo (30 seats included, more at $1 a seat up to 200 seats, 25 TB, 200 workspaces, 3,000,000 credits a month). Maximum upload sizes reach 25 GB on Starter, 50 GB on Business, and 100 GB on Enterprise. Credits meter AI operations, with overage billed at $10 per 100,000 credits. Extra storage is 1.5 cents per GB a month and extra bandwidth is 4 cents per GB. Check full plan details on the Fastio pricing page.

Evaluating document upload boundaries across leading AI platforms helps teams select the right environment for specific document processing requirements.

Anthropic Claude provides distinct attachment models across its web app and project workspaces. In standard Claude chat, users can upload up to 20 files at up to 500 MB each. In Claude Projects, individual documents are capped at 30 MB, but the total number of files is unlimited. The operational ceiling on a Claude Project is Claude's 200,000-token context window; once total project files approach that limit, users must remove documents or seek external retrieval.

OpenAI ChatGPT allows users to attach up to 10 files per prompt, with an individual file cap of 512 MB. ChatGPT processes documents using an internal code execution sandbox or vector retrieval store, subject to account-wide rolling usage caps.

Google Gemini balances smaller document caps in its web chat with larger developer file capacities. The Google Gemini web interface restricts uploads to 10 files per prompt at 100 MB per document, while the developer Files API supports files up to 2 GB with a 48-hour retention lifecycle.

The following table compares file constraints across the three major platforms:

Each platform serves distinct use cases. Google Gemini provides unmatched video analysis capacity (up to 2 GB) and an expansive 2-million-token window in Gemini 1.5 Pro. Anthropic Claude offers larger per-file document sizes in standard chat (500 MB) and structured project workspaces. However, all three platforms share the same core limitation: prompts and projects are designed for focused sessions, not enterprise-wide document archives. Offloading document repositories to a persistent, searchable workspace enables teams to use Gemini, Claude, and OpenAI models interchangeably without bumping into proprietary upload caps.