Box (@Box) header
Box (@Box) profile picture

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 https://nitter.cf/t.co/oDLXCnem9S

Joined May 2008
Photo by Box (@Box)Photo by Box (@Box)Photo by Box (@Box)Photo by Box (@Box)Photo by Box (@Box)Photo by Box (@Box)Photo by Box (@Box)Photo by Box (@Box)Photo by Box (@Box)Photo by Box (@Box)
Introducing the Box Agents API. Built on the open A2A protocol, it lets developers call Box AI agents from their own apps and AI platforms, bringing Box’s enterprise content capabilities into the agent workflows customers already use. Learn more: support.box.com/hc/en-us/art…
2
1
2
9
866
Every CIO I talk to asks the same thing about agents: what stops it from doing something dumb in our files? In 12 months, "which MCP tools is this agent allowed to call?" will be a standard line in enterprise security reviews. Our answer in @Box MCP: admins choose which tools each agent gets, Guardrails block policy-breaking writes, and every tool call is logged.
Connecting an AI agent to company files takes minutes. Controlling what it does once it’s in there is the hard part. The Box MCP server puts a governed layer between every agent and your content: • Each client, from AI assistants to coding agents and custom frameworks, runs under the user’s own Box identity and permissions • MCP Guardrails set what an agent may do, such as no deletes and no new files in externally shared folders • Classification policies decide which content an agent can reach • Every request lands in the audit trail More: blog.box.com/box-mcp-server-…
1
2
8
360
We’re bringing even more deployment options and security controls for our MCP server. With Box as the filesystem for your AI stack you can have one governed and secure platform to bring the right information across millions of files in seconds.
Connecting an AI agent to company files takes minutes. Controlling what it does once it’s in there is the hard part. The Box MCP server puts a governed layer between every agent and your content: • Each client, from AI assistants to coding agents and custom frameworks, runs under the user’s own Box identity and permissions • MCP Guardrails set what an agent may do, such as no deletes and no new files in externally shared folders • Classification policies decide which content an agent can reach • Every request lands in the audit trail More: blog.box.com/box-mcp-server-…
1
6
352
Connecting an AI agent to company files takes minutes. Controlling what it does once it’s in there is the hard part. The Box MCP server puts a governed layer between every agent and your content: • Each client, from AI assistants to coding agents and custom frameworks, runs under the user’s own Box identity and permissions • MCP Guardrails set what an agent may do, such as no deletes and no new files in externally shared folders • Classification policies decide which content an agent can reach • Every request lands in the audit trail More: blog.box.com/box-mcp-server-…
2
2
2
6
1,109
60% of people who use AI at work say their company’s AI tools aren’t keeping pace with AI. They already use it, and they can see the gap. 85% say switching models shouldn’t disrupt their work. New research on what knowledge workers actually want from enterprise AI: blog.box.com/knowledge-worke…
2
3
878
Agents are only as useful as the content they can reach. Join us at AWS re:Invent 2026 for AIM353, a hands-on builder session where you’ll build a full-stack contract compliance agent using Box MCP and @AWS in a multi-MCP architecture. Builder session · Level 300 · Nov 30 to Dec 4 · Las Vegas Register below and add AIM353 to your agenda. aws.amazon.com/events/reinve…
3
6
906
At its recent #BoxWorks conference, @Box introduced new AI tools to help contextualize and retrieve information from unstructured content stores, and find new ways to optimize process automation. @philww caught up with Ben Kus, CTO at Box, to hear more: bit.ly/4ADvj8E
3
5
426
A company brain shouldn’t just answer questions. It should learn from the work your team is already doing. @41five shares how to build that feedback loop with Box.
2
4
10
1,401
“You just have to start playing with the tools. The only way to know where your job function lands is to adopt them.” @levie in @FastCompany on why coding adoption is going vertical while other knowledge work diffuses slower, and the applied layer that bridges the two.
1
3
9
1,644
What a week in AI… I’m most excited about @Google Gemini 4 Hydrogen for our customers doing complex knowledge work every day—why? The speed and efficiency. In @Box’s evals, it completed tasks more than 2x as fast as Gemini 3.8 Flash, using about 12% fewer tokens—while improving accuracy from 74% to 79%. For customers running complex content workflows again and again, wait times and token usage add up. These gains make AI more practical for everyday work. And accuracy matters just as much—faster reports only help if the numbers are right. Excited to continue our partnership with Google! Gemini 4 Hydrogen is now available to our customers through Box AI. Read more about our evals here: blog.box.com/gemini-4-hydrog…
Replying to @GoogleDeepMind
Now rolling out for: Developers: Build with the Gemini API in @GoogleAIStudio and Android Studio, and in Google @Antigravity for paid Google AI Pro and Ultra subscribers. Enterprises: Deploy via Gemini for business and Google @Antigravity Everyone: Access it in Google Search and the @Geminiapp model dropdown. See the details → goo.gle/4zDCs8N
7
9
695
Review the process. Build the workflow. Approve improvements. In our Box Architect preview, we show how the proposed workflow becomes metadata, automation, an app and agents, with human review. Check out more on Box Architect: blog.box.com/box-unveils-arc…
1
2
7
69,991
Box retweeted
Slite Agent now answers from @meetgranola, @MiroHQ, @Box, and @mondaydotcom. Ask Slite Agent about your meetings, boards and projects, right next to your knowledge base. Everyone can connect their own account, so answers only use what each person can see. Set it up in one minute: 1. Go to Settings › Agent sources 2. Pick your source of choice 3. Click Connect and log in with your own account Using another tool with an MCP server? Add it as a custom source and Slite Agent answers from that too. Set up a Custom MCP: utm_source=x&utm_medium=social&utm_campaign=slite-sources&utm_content=slite-page
3
5
12
750
We put Gemini 4 Hydrogen from @GoogleDeepMind to work reviewing a $486,000 construction change request in Box. Gemini identifies a crane charge that was already paid and reconciles the request to $330,000 using the signed terms and project records. It also flags a time extension that lacks supporting evidence. Box AI saves a source-cited recommendation for the project owner’s review. Customers can build custom agents for their own document-heavy workflows with Box AI Studio.
2
3
6
63,646
Box retweeted
Agents need an easy, secure way to work with files. Last week at BoxWorks, @Box announced a cleaner way to give them access: Box Mount. With Box Mount, a Box folder shows up as a local directory inside an E2B sandbox. Claude reads the contract there, checks it against your legal playbook, and the review memo syncs back to Box in seconds. Box keeps the files. The agent runs in E2B and never sees your Box credentials. Check out the example: github.com/e2b-dev/e2b-cookb…
5
5
1
30
1,860
Ok this is a big deal. We’ve been testing Gemini 4 Hydrogen in early access with the Box AI Complex Work Eval, and it’s a serious jump for fast models. Hydrogen is 5 percentage points better in accuracy than Gemini 3.8 Flash. That alone would be a great boost in capability, but possibly more importantly it did the tasks in less than half the time with about 12% fewer tokens on a typical task. For this eval, we give the Box Agent Hydrogen as a model for a series of complex tasks that represent real world work with enterprise documents across finance, consulting, life sciences, and other industries. Hydrogen will be a great model for a lot of efficient but complex tasks across these fields. Gemini team is definitely cooking. Hydrogen will be rolling out in the Box AI Studio shortly so any customer can build a custom agent with it.
Introducing Gemini 4 Hydrogen, an extremely capable and fast model built to bring Gemini 4 to everyone! Available to all tiers of Google AI Studio, Gemini API, Gemini for business, the Gemini app, Google Search, and paid Google AI Pro + Ultra subscribers in Google Antigravity.
115
57
12
826
106,610
🆕@GoogleDeepMind's Gemini 4 Hydrogen is here. We ran it against Gemini 3.8 Flash on real enterprise document tasks, and it scored 79% vs 74% while finishing the work in less than half the time. → Public sector: 95% vs 69% (+26 points) → Media & entertainment: 76% vs 57% (+19 points) → Life sciences: 70% vs 56% (+14 points) → Technology: 78% vs 68% (+10 points) → Healthcare: 81% vs 76% (+5 points) It reads spreadsheets, tables and images precisely, carries multi-step calculations through to the final figure, and lands on the same answer run after run. Read the full evaluation here. 👇 blog.box.com/gemini-4-hydrog…
Meet Gemini 4 Hydrogen: built for rapid coding, multimodal synthesis, and everyday workflows. It finishes tasks 2.5x faster than 3.8 Flash and surpasses Argon on the @ArtificialAnlys Intelligence Index.
2
2
13
707,212