Helping devs build the future of intelligent, content-driven apps using @Box. Check out our docs and sign-up for a free developer account on box.dev

Based in United States
Don't miss this one next week. We are joining @daytonaio to show how to give AI agents a real computer pre-loaded with governed enterprise content. Using Box Mount, a Box folder mounts directly into a Daytona sandbox as a normal POSIX directory, kept in two-way sync. An agent reads contracts, analyzes spreadsheets, generates reports, and writes results back to Box, all inside a disposable, permission-scoped environment that leaves no residue. September 29 at 9AM PT. Register here. events.zoom.us/ev/Amkp36LSSk…
6
5
15
1,456
"We still need a layer that can bring the power of those models into the real workflows in the enterprise. And this is the opportunity for the applied AI layer." @levie on @CNBC on why model progress alone is not enough, and why the companies bridging AI breakthroughs into real enterprise workflows are where the next years of value get created.
9
6
41
106,758
Reliable extraction requires testing and refinement. The new Box Extract agent builder makes it easier to define extraction instructions, test them against your content, and refine results, all in one streamlined experience. See it in action.👇
1
6
21
59,877
We gave Meta’s Muse Spark 1.3 a product launch to rescue: 7,000 units due in one week after a supplier shutdown. In this Box AI demo, one prompt sends it through 24 documents to reconcile conflicting inventory records and reject an incompatible substitute. It generates the lowest-cost feasible recovery plan: $103,700 to meet every shipment deadline, with a cited daily schedule and the approvals needed before anyone buys materials or books freight. Muse Spark 1.3 is now available in Box AI.
4
9
31
364,050
Muse Spark 1.3 is now available in Box AI. What our evaluation of Muse Spark 1.3 found: → 13% accuracy lead over comparably priced models on the Box Complex Work Eval → 42% faster than Muse Spark 1.2 using roughly a third fewer tokens → Strongest gains in Financial Services, Technology, and Energy
8
8
48
384,455
Box is coming to @Muse! Meta's personal AI agent will now be able to work with your content in Box, so the files, documents, and knowledge your team depends on are accessible right where the work is already happening.
2
4
17
1,535
We gave GPT-6 Sol 14 documents to check a household’s wealth management fees. It found $5,125 in overcharges. In this Box Agent demo, @ChatGPT's Sol applies a signed fee amendment, catches an asset that should have been excluded from billing, and reconciles a $1,000 credit already posted. That leaves $4,125 in proposed corrections for human review. Sol also confirms that a separate trust account’s prorated fee was correct, then produces a cited, account-by-account review for Billing Operations and Compliance. Explore Box AI Studio to build custom agents for your own document-heavy financial services workflows.
4
22
408,784
On September 29, join us for a session with @daytonaio on how to give AI agents a real computer and the enterprise content to work on. What you'll learn: → How to think about the right use cases for sandboxes vs. legacy infrastructure → How to configure and operate Daytona composable computers → How to connect enterprise content to your agents using Box Mount If you're building agents that need to do real work on real content, you won't want to miss this. Register here. events.zoom.us/ev/Amkp36LSSk…
1
7
17
5,002
We put @claudeai Opus 5.5 to work on a $19 million reinsurance recovery where the right answer was buried across 14 documents in Box. In this Box Agent demo, Opus applies an amended occurrence window, reconciles claim payments with other recoveries, and separates estimated recovery from what can actually be called today. The records support $9.8 million in estimated recovery, but only a $3 million additional cash call. The agent generates a cited review for the insurer’s reinsurance team, with the outstanding approvals identified. Claude Opus 5.5 is now available in Box AI.
8
19
331,441
Process lives in @ServiceNow. Content lives in Box. Today we are connecting the two natively. Box for ServiceNow is now available on the ServiceNow Store. A Box panel sits directly on the ServiceNow record so teams can navigate folders, preview files, upload, and share content without switching systems. Every governance policy stays in place because your files never leave Box.
3
7
1,340
The context developers need to build better applications has always lived in Box. Now it can securely travel with them into @vercel. With Vercel Connect, developers build agents that find, organize, and act on Box content without recreating its governance model. In our demo, a Customer Intake Agent built on Vercel opens a folder of customer paperwork, uses Box AI to identify each document, and files it into the right customer folder with a readable name. No stored API keys. No manual sorting. Box stays the source of truth. Watch here. 👇
4
3
52
72,379
We evaluated @claudeai Opus 5.5 on Box's Complex Work Eval. The standout finding is efficiency. → 30% faster than Opus 5 → A third of the tokens → Answers 42% shorter with the findings intact What disappeared was restatement and preamble. Opus 5.5 simply led with the conclusions. Accuracy held across Consumer Products, Financial Services, and Technology. Claude Opus 5.5 is coming to Box AI soon.
Introducing Claude Opus 5.5, the first model in our new Claude 5.5 family. It performs at the level of Claude Fable 5.1 for most tasks, and costs 40% less to run than Opus 5.
6
4
34
827,129
BoxWorks 2026 is where you get the inside track on where Box is headed next. → Keynotes from Box Co-founder and CEO Aaron Levie, @nvidia Founder and CEO Jensen Huang, and @intel CEO Lip-Bu Tan → A live walkthrough of Box's latest innovations → First look at the product roadmap from Box leadership What does your team need most from enterprise AI right now? Save your spot: boxworks.box.com/series/boxw…
5
15
2,980
New this year, join us for the first ever Dev Summit at BoxWorks. If you’re building agentic and content-driven enterprise applications and workflows you won’t want to miss this! 2 days of technical sessions and hands-on workshops covering everything from the latest in agentic sandboxes to file systems for agents. Speakers from @nvidia, @OpenAI, and Box and more coming soon! November 5-6 in San Francisco. Register here: boxworks.box.com/series/boxw…
2
13
3,258
A claims system collects evidence into a case folder. Someone has to decide whether that evidence is complete and acceptable. That decision has to be recorded. This demo routes evidence files through Box Automate. Your app triggers the workflow, which can be fully customized. Box can assigns the approval task, extract metadata, send evidence for e-signature, and more. Box allows for a smooth process, and you define the process. Demo.👇
1
7
1,679
We put Grok 4.7 from @SpaceXAI to work on a $2 million insurance claim where one deductible error alone changes the calculation by $143,000. In this Box Agent preview, @grok reconciles the claim against the policy and supporting records. It catches an $82,000 duplicate invoice and a missing $64,000 supplier credit. It also explains why the Business Income waiting period doesn’t apply to Extra Expense. The result is a cited claims review for the adjuster, with final coverage and payment decisions left to the insurer. Explore Box AI Studio to build custom agents for your own document-heavy workflows.
Grok 4.7 is here. It's a notable improvement over Grok 4.6 at the same price and speed.
78
186
1,537
1,866,345
🆕@typesafeai dropped Jev this week, a new frontier model optimized for decisions. We put it to work on an incident triage workflow in Box. Jev doesn't generate text. You send it state plus typed questions and it returns booleans with probabilities, enums, and scores in a single pass, so there's no output to parse and no schema to keep in line. The demo uses Jev to classify and route incidents. It pulls the incident report from Box, builds the state, asks whether it's customer-facing and how severe it is, moves the file into Escalate, Monitor, or Review folders, and sets a metadata template instance with the result. The answers comes back with a confidence estimate, so anything below threshold goes to Review for a human instead of routing itself. Check it out.👇
6
14
58
184,766
Box and @OpenAI are partnering to bring governed enterprise content to AI-powered legal work with Astra for Law and @ChatGPT. Legal skills built on Box and ChatGPT help law firms turn their content into practical, secure workflows without weakening the controls around it. In this demo, a legal professional uses the new Outside Counsel Guidelines skill to review an incoming set of OCGs directly from Box. ChatGPT flags associated risks and leaves comments for human review, all without the content leaving Box.
Astra for Law: Frontier intelligence built for your practice. A new offering powered by GPT-6 Astra with tools, settings, and context to support the expertise and judgment of lawyers and legal technology firms.
6
6
27
959,129
🚀@vercel's’ fx is a tiny open-source coding agent built to take action directly in a developer’s workspace. This demo extends that idea by giving fx access to the security policies, requirements, and release docs stored in Box. Using the Box CLI, fx retrieves an approved policy, compares it with local code, edits the files, adds missing test coverage, and prepares release evidence without being told which lines to change. It asks for human approval before uploading anything, then verifies every file after it lands in Box. Git stays where the code lives. Box stays where the approved docs and release proof live. fx connects the work across both. Demo below ↓
2
4
35
305,445
Going to @WeAreDevs in San Jose? 🎮 Come check out AI Trivia and Game Night, Wed Sept 23, 5-7pm at Guildhouse. 2 min walk from the convention center. Food, drinks, games, and the DevOps + SRE crowd. Hosted by @pagerduty, Box, @arizeai, and @EntireHQ. RSVP 👉 luma.com/prtcc7pw?tk=byxXsF&…
1
4
10
2,523
The applied layer is where AI creates enterprise value. @levie joined @sonyatweetybird to discuss why the most intelligent model in the world still has to connect to legacy systems, wait on humans in the loop, and work within workflows that were never designed for AI.
3
7
54
842,144
Claude’s new Slides experience can turn deal materials in Box into an editable credit committee briefing. In our demo, @claudeai reviews a borrower’s deal workspace, creates a five-slide presentation with financial visuals and source references, and flags conflicting information for review. We leave a comment directly on a slide, Claude revises it, and the finished deck is exported as PowerPoint and uploaded back to Box through MCP for the credit team to review. Connect Box to Claude to create and refine presentations using your enterprise content.
Replying to @claudeai
You can also now make decks, docs, and designs in your conversation. Draft the one-pager in Claude Docs, turn it into a deck with Claude Slides, and mock up a matching visual in Claude Design, all from one place.
4
3
37
735,223
Box + @Snowflake turn financial documents into queryable history, without losing the path back to the source. Most financial research workflows follow the same pattern: extract details from documents, write a summary, copy numbers into another system, and hope everything can be retraced when someone asks where a figure came from. This Python demo closes that loop. Box Extract transforms PDFs into structured financial data while preserving confidence scores and page references. Box Doc Gen produces a readable, sourced memo. Snowflake stores common fields alongside the complete JSON payload, creating queryable records that can be compared across quarters or companies. The result is unstructured documents become useful data while every number remains connected to its source.
4
4
16
161,843
Structured extraction starts with defining the right schema. With Box Extract, AI can analyze representative documents and generate metadata fields, field types, and extraction instructions, giving teams a faster starting point for extracting structured data at scale. See how it works.👇
4
4
20
87,557
Enterprises are sitting on massive amounts of unstructured data that they've never been able to fully tap into before AI. The big opportunity now for organizations is to finally be able to understand and and get the value from this information at scale with AI agents. @levie spoke with @sonyatweetybird on why agents that can read, extract, and act on unstructured content are how real enterprise value is created.
Well.. @levie and I filmed this episode of Training Data a week or two ago, when the “current thing” was Doug Leone’s novacaine root canals instead of pacing the frontier… Simpler times! But Aaron’s advice on reinventing yourself and your company for AI is timeless. Aaron founded @Box 20 years ago. It sits on hundreds of billions of enterprise files, and he's bet the company on agents that can read every one of them. He's also one of the most wired-in people in AI, on every cap table and, by his own admission, 95% Twitter-educated. He’s the rare CEO who can straddle both the internet AND has the ear of CIOs. His core argument: (1) the gap between what a model can do and what an enterprise workflow actually needs is vast, and closing it is a lot of software; (2) diffusion of AI outside of coding will take far longer than Silicon Valley thinks, and that slowness is exactly where the applied layer's value comes from. The conversation covers: — why application companies are the hottest neolabs, and why the LLM-wrapper thesis is finally working — the fox-guarding-the-henhouse problem with letting model providers route your tokens — work slop, and why we accept AI-written code but flinch at AI-written decks — how Box built its agentic harness and why it beats raw API access on accuracy and latency — the open-weights paradox: closed labs and open models both growing exponentially at once — what continual learning has to solve before it works for a lawyer with five matters and a Chinese wall — why 90% of enterprise tokens in five years will come from tasks no human kicked off — the mandate for founders right now: whoever gets it to the customer wins 0:00 – Introduction 1:55 – Are application companies the hottest neolabs? 6:56 – Will the labs move up the stack? 12:34 – Box and betting the company on AI 16:50 – Hero use cases: reading a million contracts and long-running agents 18:42 – Work slop: why AI code is embraced but AI content isn't 24:08 – Building Box's agentic harness and the evals that matter 27:23 – The state of the model race 29:25 – Open-weight model adoption in the enterprise 32:34 – Memory, continual learning, and what belongs in the weights 37:29 – Box Labs and systems of record in a world of agents 44:55 – Will chat be the dominant UI for enterprise AI? 48:00 – Why coding diffused fast and the rest of knowledge work hasn't 54:31 – Staying wired in, making a company AI-first, and what it takes to win
6
1
22
202,089
Box is connected with @ChatGPT for Financial Services, so firms can analyze financial data alongside their own research and investment materials. In this demo, OpenAI reviews 14 documents in Box and finds that 20% reported growth translates to 8% organic growth. It refreshes the valuation from $56 to $47.92 per share and drafts a source-cited investment committee update for human review. We're excited to be a launch partner with @OpenAI.
1
5
44
411,138
AI tools will keep changing, but content policy needs to travel with the content. See how Box Shield applies classification-based access controls to what AI agents can actually read, not just what they can download. 👇
4
10
101,060
🆕 @OpenAIDevs Agents API gives developers a hosted sandbox where agents run tools, coordinate work, and handle complex tasks. Box Mount brings enterprise content directly into that sandbox as files the agent can read, reason across, and produce new work from, using normal shell commands and file paths while Box Mount handles two-way sync automatically. In our demo, a lead agent mounts a deal room from Box, reads five source files, launches three specialist agents in parallel, and writes four reports back through Box Mount with an initial verdict. When a customer revises the MSA in Box, Box Mount syncs the new version into the same sandbox and the agents reassess automatically, refreshing all four reports while preserving the revised MSA as version two in Box. The result is a shared workspace where people and agents work on the same governed Box content without building custom file-transfer logic, with permissions, versions, and audit history preserved throughout. Box Mount is in private preview. Watch👇
Go from idea to a working agent faster with the Agents API. Build and run cloud agents with the Codex harness, fully managed by OpenAI. We handle orchestration, long-running sessions, and context management. You focus on what makes your agent unique. Available in public beta.
19
11
83
230,250
We've partnered with OpenAI to bring enterprise content directly into @ChatGPT. In our demo, someone inherits a lease, finds it in their Box folder from ChatGPT, gets certain sections flagged with the exact language visible in the preview, and shares it with legal without switching tools or moving the file. This is Box as the file system for AI.👇
8
5
24
54,893
OpenAI and Box are bringing your enterprise content directly into @ChatGPT. As the file system for AI, Box provides headless access to enterprise context in ChatGPT while keeping the content managed, governed and protected by your existing Box permissions. Demo ↓
And starting today, @Dropbox, @Box, and Sharepoint are now natively integrated into the ChatGPT Library across both ChatGPT and ChatGPT Work. Rolling out this week to paid users.
19
12
151
3,477,882
Collateral disputes in derivatives trading often come down to one question: is the requesting firm using the right numbers? In this @claudeai Cowork demo, we use the Box MCP server and a reusable skill to review ISDA agreements, the contracts governing derivatives trading, and check whether firms are requesting the right amount of collateral. Claude shows how contract amendments and credit-rating changes affect those amounts, with a cited exception report and structured metadata saved back to Box for human review. Watch here.👇
3
8
1,757
We’re introducing Box Mount. Mount a Box folder into a sandbox like @e2b and keep it in two-way sync. The agent reads and writes Box content with standard file operations as if they were local. Every operation flows through the Box API so governance applies automatically, no custom policy code required. 2 commands to get started: agent-mount config agent-mount mount "/mnt/box" "<box-id>" Developer Preview is now open.
5
8
48
56,155
We put GPT-6 Astra from @OpenAI to work on some of the hardest reasoning in enterprise knowledge work, untangling conflicting records and different approval requirements across a complex loan amendment. In this Box Agent demo, Astra reviews 23 documents to determine which lenders can vote on a loan amendment. It corrects the preliminary tally and shows why the base amendment has enough consent while changes to pricing, maturity, and collateral do not. It also identifies a covenant breach that an earlier waiver doesn’t cover and saves a source-cited readiness report back to Box for human review. Customers can build custom agents for their own document-heavy lending workflows with Box AI Studio.
3
4
23
472,227
GPT-6 Astra from @OpenAI is built for long-horizon analytical work. We evaluated it against GPT 5.6 Sol on Box's Complex Work Eval, real enterprise document tasks where you pull the source files, reason across spreadsheets, contracts, and filings, and produce the actual deliverable: → Overall: 77% vs 74% → Technology: 77% vs 62% → Media and Entertainment: 77% vs 62% → Legal: 72% vs 64% Frontier performance on the most demanding workflows, where documents disagree, the deliverable requires a figure nobody handed you, and the whole job is reconciling what the numbers say with what the narrative claims. GPT-6 Astra is coming soon to Box AI Studio.
This is GPT-6 Astra. Anything you can do on a computer, Astra can do for you. Fast.
4
10
41
868,093
OpenClaw 2.0 is out and /loop is now part of the stable release. Connected to Box, here's what that looks like. One prompt connects @openclaw to an M&A deal room via the Box CLI and starts a recurring check. When a signed customer renewal lands in the folder, OpenClaw catches it on the next pass, connects it to the existing risk register, updates the relevant deal context, and saves the update as a new version of the same file in Box, with the agent identity preserved throughout. Start it with one prompt. Stop it with /loop stop. OpenClaw keeps up with the work as it changes, and Box stays the source of truth.
4
2
12
2,749
Introducing Box Agent in Apps. A multi-step reasoning agent built directly into Box Apps that breaks complex goals into subtasks, searches your content, reads documents, and delivers a traceable answer without you opening a single file. Compliance questions that would take hours of manual contract review now take minutes, with everything staying inside Box under your existing governance policies and access controls.
4
2
10
1,712
🆕@claudeai Fable 5.1 can reason through capital-markets operations where a trade can match economically and still not be ready for settlement. In this Box Agent demo, it reconciles five institutional equity trades, catches a 500-share allocation shortfall, an $18,000 pricing error, an incorrect settlement date, and an outdated settlement instruction, then saves a cited resolution pack to Box. Customers can use Fable 5.1 in Box AI Studio to build custom agents for document-heavy financial services workflows.
5
4
54
562,189
Anthropic’s @Claude Fable 5.1 is built for multi-step analytical enterprise work. We ran it through Box's Complex Work Eval against Fable 5 : → Overall: 72% vs Fable 5's 65% (+7pp) → Financial Services: +19pp over Fable 5 → Technology: +37pp over Fable 5 → Public Sector: +16pp over Fable 5 → Data Analysis: +12pp over Fable 5 A clear step up on the quantitative, multi-step work that drives real decisions. Fable 5.1 is 23% faster and uses 25% fewer total tokens than Fable 5. More accurate and more efficient at the same time.
We’re introducing Claude Fable 5.1 and Claude Mythos 5.1. They're the world’s most advanced models for coding and knowledge work.
15
8
78
786,711
Box has signed @OpenAI's call for collective action on cyber defense, alongside more than 150 other industry leaders. Our contribution starts at the content layer, where the most sensitive enterprise data lives and where the most important security decisions get made. AI-enabled attacks are becoming more sophisticated. The window to act is now. Status quo security will not be enough. Read the full letter here. 👇 openai.com/collective-cyberd…
4
4
14
319,533
Box Drive now connects directly to @cursor_ai from your desktop. Right-click any file or folder in Box Drive and open it directly in the Cursor desktop app, giving you instant access to your Box content without downloading files or switching between tools. In this demo, a developer right-clicks a PRD stored in Box Drive, opens it in Cursor, and asks it to build an app based on the specs. Cursor reads the full requirements document and generates the results, all without the spec ever leaving Box. Your product specs and your development environment finally in the same flow.
3
6
15
150,455
Claudeforce gives @claudeai access to the context and actions in @salesforce. But the full story of your business doesn’t live in CRM alone. With Box MCP, Claude can also securely tap into the contracts, proposals, policies, and other enterprise content behind the work. Box + Claudeforce: bringing governed enterprise content to every AI agent.
Welcome Claudeforce. ❤️
7
2
57
304,458
We reported strong Q2 earnings as companies accelerate efforts to give AI agents access to their critical corporate information and unstructured data. @levie spoke with @CNBC about what we're seeing across every major market and industry. From financial services, life sciences, government, media and technology. "Every enterprise on the planet is sitting on an incredible wealth of information."
19
9
70
442,464
Today, Box reported our second quarter fiscal 2027 financial results. Q2 was another strong quarter, driven by record Q2 bookings and increasing Box AI adoption. As a result, revenue, billings, and operating margin all exceeded expectations. See the full earnings announcement here: businesswire.com/news/home/2…
3
3
13
1,897
Most teams governing AI agents focus on what they can access. The harder problem is what they can do with it. Box MCP Server admin controls let admins govern tool access enterprise-wide or per AI integration, set read-only or read-and-write permissions, and configure down to the individual tool level. When a tool is disabled, agents can no longer discover or invoke it. Enforced at runtime, not just at setup.
7
6
32
475,070
Two days out. 📆 Your agent can browse the web. But can it log in, grab a file from behind a portal, and put it somewhere your team actually works from? @chriskim_dev and @kylejeong are showing you how on August 26. @browserbase for the browser. Box for the file system. Your agent connects them. Ready to ship real workflows and not just prototypes? Come build with us on Wednesday. Register here: luma.com/8vn867oe
1
8
1,780
The Box connector, powered by the Box MCP server, is now live in @GeminiApp Enterprise. In our demo, Gemini reads six vendor contract PDFs, resolves a quiet amendment, and writes structured metadata back onto every file using Box Extract. A second prompt builds an interactive HTML dashboard - to not only see all of that metadata in one consolidated view, but also highlight key details - such as $760K in contract value auto-renewing in the next 90 days or another contract that is still cancellable for five more days - so that humans can take action on those insights. Unstructured PDFs to structured metadata to an actionable dashboard. This is what it means to turn unstructured data into actionable business insights.
3
4
40
1,056,618
Grok @bot is now connected to Box through the Box MCP server. Personal AI built for complex knowledge work, with secure, governed access to your enterprise content from day one.
We're making Grok Bot more widely available. All SuperGrok Plus, Cursor Pro+, and Cursor Teams subscribers now have access. We're also offering a free trial with limited usage for all other users.
1
2
37
6,200