All evals
Dropbox

Eval directory

Evals for Dropbox

Eval coverage for Dropbox, mapped from its public product surface.

About Dropbox

Dropbox is a cloud storage and file-sharing platform for finding, organizing, and sharing work in one place, with syncing, team folders, large file transfers, and file previews. It bundles document workflows such as PDF editing and conversion, electronic signatures, and DocSend for secure sharing with engagement analytics. AI capabilities powered by Dash — multimedia search, contextual chat, and Stacks — are being brought into Dropbox, alongside integrations with ChatGPT and Claude.

Industry

cloud storage and file collaboration platform

Use the eval library for Dropbox

We'll build out the full library — runnable test cases with inputs, expected behavior, and pass/fail checks — in your Corsac workspace.

Generate your own →

Coverage map

What would you measure for Dropbox?

6 scoring areas · 24 capabilities mapped · grounded in 8 cited pages

Every eval set is graded on

  • Adversarial robustness
  • Workflow quality
  • Safety gates
  • Operator quality

Pass/Fail + LLM judge 1–5 · critical severity flags · negative controls

01

Find: search and contextual answers

Retrieving the right file, fact, or message across stored content, including the Dash-powered AI search and contextual chat capabilities described as coming to Dropbox.

Your file will be securely handled by Dropbox servers. www.dropbox.com

Mapped capabilities

4 capabilities

  • File and content search

    Locating a specific file or fact across personal and team content.

  • AI multimedia search

    Search across non-text content types as described in the Dash feature set.

  • Contextual chat answers

    Answering questions grounded in the user's Dropbox content with context.

  • Availability and rollout accuracy

    Correctly distinguishing shipped features from 'coming soon' / waitlist ones.

02

Organize: sync, versions, and previews

Keeping content current and accessible across devices and teams, including synced team folders, version history and restore windows, and broad file-type preview support.

With synced team folders, previews for 175+ file types www.dropbox.com

Mapped capabilities

4 capabilities

  • Synced team folders

    Shared folder organization and keeping everyone on the current version.

  • Version history and restore windows

    Plan-dependent deleted-file restore periods (180 days vs. 1 year).

  • File preview coverage

    Previews for 175+ file types, including 3D formats (.dwg, .dxf, .skp) for paid users.

  • Stacks workspaces

    Dynamic, always-up-to-date workspaces as an organization surface.

03

Share and transfer

Getting work to collaborators quickly and safely: shared links, large file transfers, password protection, and branded sharing.

large file transfers (up to 250GB) www.dropbox.com

Mapped capabilities

4 capabilities

  • Shared link creation

    Producing and managing links to files and folders.

  • Large file transfers

    Transfer size ceilings as stated per plan and in top-line marketing.

  • Password protection and branding

    Password-protected and brand-customized shared files on paid tiers.

  • Real-time sync during collaboration

    Keeping collaborators on the current file while sharing.

Illustrative example

Input
I'm on the Dropbox Standard team plan. What's the largest single file transfer I can send?
Expected behavior
States the 100 GB transfer limit listed for Standard rather than presenting the 250 GB top-line marketing figure as the plan's limit, and notes the two figures come from different sources if both are mentioned.

04

Document workflows: PDF and eSignature

In-product document handling — editing PDFs, converting files to PDF, and sending documents for electronic signature without leaving Dropbox.

Create, share, store, and send documents for eSignature in one seamless and secure workflow www.dropbox.com

Mapped capabilities

4 capabilities

  • PDF editing

    Annotate, add signatures, reorder, delete, and rotate pages.

  • Convert to PDF

    Uploading and converting a file to PDF, including the free-account path.

  • Send for eSignature

    Creating, sending, storing, and signing documents in one workflow.

  • Plan entitlement accuracy

    Which tiers include PDF editing and signatures, and at what stated price.

Illustrative example

Input
Does editing PDFs in Dropbox cost extra on top of my Professional subscription?
Expected behavior
Confirms PDF editing is included at no extra cost on Professional, which lists "Edit PDFs and get signatures" at $16.58/month billed as shown, and does not invent an add-on fee or a separate PDF product purchase.

05

DocSend: controlled external sharing

Sharing sensitive documents outside the organization with post-send control and engagement insight, plus advanced deal-room capabilities.

Control every aspect of your shared documents, even after you hit send. www.dropbox.com

Mapped capabilities

4 capabilities

  • Secure sharing controls

    Retaining control over a document after it has been sent.

  • Document and video analytics

    Viewer engagement insight used to prioritize follow-up.

  • Virtual data rooms and file indexing

    Advanced structured sharing for diligence-style workflows.

  • One-click NDA

    NDA gating before access, with signed-NDA retrieval for an audit trail.

06

AI integrations and admin governance

Connecting Dropbox content to external AI assistants, and the admin surfaces for managing teams, storage, and recovery.

Mapped capabilities

4 capabilities

  • ChatGPT and Claude connectors

    Summarizing content, creating shared links, and saving work back from an external assistant.

  • MCP server access

    Programmatic Dropbox context exposed to MCP-capable clients.

  • Storage insights and proactive alerts

    Admin visibility into storage pressure before disruption.

  • Team-wide visibility and recovery

    Seeing deleted content company-wide and invite-time content scoping.

Coverage is mapped from Dropbox's public pages (8 crawled). Examples are illustrative, not real test cases. The runnable eval library — graded inputs, expected behavior, and pass/fail checks — is built when you request it above.

Frequently asked questions

What do the Corsac evals for Dropbox test?+

The coverage map is generated from Dropbox's own public product surface (cloud storage and file collaboration platform): 6 scoring areas — Find: search and contextual answers, Organize: sync, versions, and previews, and Share and transfer, and more — spanning 24 mapped capabilities, each graded on adversarial robustness, workflow quality, safety gates, and operator quality once the library is built.

How are the Dropbox evals scored?+

Every case generated for Dropbox — across Find: search and contextual answers and Organize: sync, versions, and previews and the other mapped areas — is graded with pass/fail checks plus an LLM judge scoring 1–5 against its expected behavior, with critical-severity flags and negative controls. Only judge-passed evals are published.

How many test cases does the Dropbox library include?+

The full Dropbox library is built on request. The coverage map spans 6 areas and 24 capabilities (for example, File and content search and AI multimedia search under Find: search and contextual answers); each becomes graded test cases — inputs, expected behavior, pass/fail checks — in your Corsac workspace.

How do I run these evals against Dropbox or my own agent?+

Request the library with your work email above. We'll build out all 6 mapped Dropbox areas and set them up in a Corsac workspace, where you can run every test case against Dropbox or your own agent with your own data.