Perplexity AI File Uploading: Size Limits, Supported Formats, Plan Differences, and Workflow Strategies for Late 2025/2026

Perplexity AI has steadily expanded its file-uploading capabilities, enabling users to analyze, search, and query across documents, spreadsheets, images, audio, and video.This article presents a comprehensive, up-to-date guide to Perplexity’s file upload limits, supported file types, plan-specific features, new workflow options, and strategic recommendations for getting the most from this evolving platform as of late 2025/2026.

····················

Perplexity AI allows file uploads across all major plans, with increasing size and retention limits at higher tiers.

All users—including free, Plus, and Pro tiers—can upload files for analysis, question-answering, and document search.The standard web-app file size limit is 40 MB per file for both Free and Plus plans, supporting up to ten files per prompt with a retention window of 30 days.

Pro users benefit from an increased limit of 50 MB per file (available via Labs or API) and 90-day retention, while Enterprise Pro accounts unlock up to 1 GB per file and a one-year retention window, suitable for organizational data pipelines or large research archives.API attachments for Sonar or PPLX can accept 50 MB base-64 payloads and also permit external URLs to bypass the hard cap for large datasets.

···············Perplexity AI File Upload Limits by Plan

Plan	Max File Size	Files per Prompt	Retention
Free / Plus	40 MB	10	30 days
Pro	50 MB	10	90 days
Enterprise Pro	1 GB	20	1 year
API (Sonar/PPLX)	50 MB • ext. URL	—	User-managed

····················

A wide range of file formats are supported, including documents, spreadsheets, images, audio, and video.

Perplexity’s file engine parses a broad array of formats:

• Documents: PDF, DOC, DOCX, TXT, RTF, ODT

• Spreadsheets: XLSX, CSV (≤ 25 MB recommended for best results)

• Presentations: PPTX

• Text: Markdown, JSON, HTML

• Images: PNG, JPEG (used for vision analysis)

• Audio/Video: MP3, WAV, OGG, FLAC, MP4, MOV, AVI, WEBM (up to 40 MB)

Unsupported or binary files, macro-enabled spreadsheets, and unknown extensions are rejected automatically.

Audio and video uploads can now be transcribed into searchable text (as of November 2025), while images inside video frames are not yet indexed for search (feature scheduled for 2026).

···············Supported File Formats (December 2025)

Type	Formats	Notes
Documents	PDF, DOC, DOCX, TXT, RTF, ODT	Up to 40–50 MB
Spreadsheets	XLSX, CSV	≤ 25 MB for best parsing
Presentations	PPTX	Standard PowerPoint
Images	PNG, JPEG	Used for visual analysis
Audio	MP3, WAV, OGG, FLAC	Up to 40 MB, transcribed
Video	MP4, MOV, AVI, WEBM	Up to 40 MB, transcription only

····················

Perplexity’s handling of uploaded files includes on-demand parsing, semantic Q&A, and retention-based retrieval.

Uploaded files are cached for the retention period based on plan, allowing users to reference, search, and query multiple documents over time.Text and metadata are indexed to support document-level search and direct question-answering.PDFs, Office documents, and other rich-text files can be queried using the document_search tool, which semantically analyzes the content for contextual responses.Large CSVs and spreadsheets may only be partially parsed—community experience suggests the free tier reliably analyzes the first ~100 kB or a few hundred rows, so large datasets should be split for full coverage.Developers can use the API to attach files in Sonar calls, retrieving embeddings and extracted text for custom applications.

···············File Handling and Retrieval Workflows

Step	Action	Purpose
Upload	Add files via UI or API	Store and index
Parse	System reads content	Create embeddings, extract text
Query	Ask questions by file or section	Semantic Q&A
Retain	Cached for 30–365 days	Enables follow-up queries
Export	Download extracted data	Further analysis

····················

Recent updates add larger file caps, audio/video transcription, and more flexible multi-file analysis.

September 2025 saw Pro file size caps bumped from 40 MB to 50 MB, offering more headroom for dense PDFs or media.October’s update enabled simultaneous upload and cross-document Q&A with up to ten files per prompt, even for free-tier users.November’s Labs launch brought audio/video transcription—converting AV files up to 40 MB into searchable text or semantic summaries.December’s API roadmap confirmed that frame-level vision analysis for videos is in development for release in 2026.

···············2025 Feature Update Timeline

Month	Feature	Benefit
Sep	Pro cap to 50 MB	Larger files
Oct	10 files per prompt	Cross-file analysis
Nov	AV transcription	Audio/video to text
Dec	API vision roadmap	Deeper video search planned

····················

Known limits, workarounds, and best practices for file-based workflows in Perplexity.

Large CSVs are best split into ≤ 1 MB segments for reliable parsing; use descriptive filenames and reference them in your prompts.Compressed or macro-free PDFs upload fastest and parse most reliably.Images and AV files are limited to visual/text extraction—embedded images in video frames are not yet indexed.For long-term research or organizational use, Pro and Enterprise tiers are recommended for increased file caps and extended retention.API attachments with external URLs can circumvent UI caps but may require authentication and additional setup.

···············Workflow Optimization Tips

Tip	Reason
Compress large PDFs	Avoid upload failures
Split big spreadsheets	Maximize coverage and parsing accuracy
Use descriptive filenames	Easier multi-file referencing
Upgrade for large projects	Unlocks higher limits, longer retention
Leverage AV transcription	Multi-modal research, interviews, lectures

····················

Perplexity AI’s file uploading makes it a strong platform for research, Q&A, data extraction, and multimodal analysis.

With growing support for large files, spreadsheets, audio/video transcription, and structured search across multiple uploads, Perplexity now rivals leading AI assistants for document-centric workflows.The combination of semantic Q&A, flexible upload policies, and API-driven integrations allows users—from students to enterprise teams—to build custom knowledge management and automation pipelines.Awareness of size caps, plan features, and parsing quirks is essential for extracting maximum value from the platform as it continues to evolve.

··········

FOLLOW US FOR MORE

··········

··········

DATA STUDIOS

··········

[datastudios.org]