Grep 2.0: A Second Brain For Serious Work
Grep 2.0 is here and it's the biggest release yet, giving every user a second brain in the cloud with expertise, memory, and execution. Plus continuous monitoring, usage-based billing, and our first case study.
Today we're launching Grep 2.0 - the biggest release yet. At the center is Brain, a new AI architecture that understands you and your work, and turns research into real deliverables - slides, spreadsheets, dashboards, full apps, in an all new chat interface that helps you get serious work done.
The Problem Grep Brain Solves
Serious work where high-stakes decisions are made require three things: domain expertise, understanding, and execution.
A compliance officer screening a counterparty needs to know which sanctions databases to check and how to trace beneficial ownership. That's expertise. They need to remember that this counterparty surfaced six months ago in a different context. That's understanding. And they need a documented report with source citations their team can review - not a chat transcript to rewrite. That's execution.
Today's AI research tools deliver none of these reliably. They search the web, summarise what they find, and forget everything when you close the tab. Every session starts from zero. The result: professionals spend hours rebuilding context, verifying sources, and reformatting output into something usable.
Grep Brain changes that.
Introducing Your Second Brain
Brain is a new AI architecture that gives every user a dedicated computer in the cloud - with its own memory, workspace, and team of domain experts. It delivers all three things high-stakes work demands: Understanding, expertise and execution.
It's not a chatbot with a better prompt or an AI agent to automate a workflow. It's a system modeled on how your brain actually works - short term and long term memory, routing requests based on context and handing off execution to more specialized agents.
Understanding You And Your Work
A senior analyst is more useful than a new hire because they already understand your business, your priorities, and the way you work. Brain starts there.
The first time you open Grep, Brain runs a short onboarding interview - your role, your industry, the kinds of decisions you make, the people and companies you watch, and the formats you ship in. That profile becomes the lens every future query is run through.
Ask "screen this counterparty" and Brain knows whether you mean OFAC compliance, vendor due diligence, or M&A target verification - because it knows what you do. The same words mean different things at a hedge fund, a bank, and a venture firm. Brain doesn't make you spell that out every time.
Brain also combines persistent files, long-term memory, and deep research into a single environment that understands your work. The competitive analysis you ran in January is there in April. The due diligence from last quarter informs this quarter's follow-up. You never start from zero.
Every session makes the next one better. Brain surfaces relevant prior work, connects dots across projects, and picks up threads without you re-explaining the background. The difference between a tool that remembers and one that doesn't is the difference between a new hire and a senior analyst.
Expertise: Over 250 Specialized Skills and Tools
Grep gives you instant expertise by drawing on more than 250 codified research skills and over 100 specialised tools - structured methodologies that encode how professionals actually work. How to verify a beneficial owner across jurisdictions. How to assess sanctions risk against OFAC, EU, and UN lists. How to evaluate a patent landscape.
Each skill knows where to look, what to verify, and how to structure findings. Multiple experts work in parallel when questions span domains. They pull from 90+ trusted primary sources: court records, patent registries, sanctions databases, corporate filings, academic papers. Not search engine results. The actual documents.
Execution: Research To Results In Minutes
Most research tools give you text and call it done. Brain builds deliverables.
Board meeting prep becomes a slide deck. Competitive analysis becomes a sortable spreadsheet. Company investigations become documented reports with source citations. Behind this is GREP's agent harness - a sandboxed environment where Brain writes code, builds applications, and deploys artifacts.
Slides (html and pptx). Spreadsheets (xls). Reports. Dashboards. Live apps. First-class outputs, not afterthoughts. All you have to do is ask for it and Grep will deliver! Because the value of research isn't in knowing things - it's in what you do with what you know.
Monitoring: Research That Never Sleeps
A research report is a snapshot. The world keeps moving.
Brain can keep watching any query you've run - on a schedule you set, or only when something material changes. Quarterly competitive analysis refreshes itself. A counterparty under sanctions review pings you the moment new filings land. A patent landscape stays current without anyone re-running it.
You set the cadence; Brain watches the underlying sources, re-runs the relevant skills in the background, and surfaces only what's new. No daily noise. No stale dashboards. The research keeps working between sessions, so you spend your time on the changes that matter - not on re-running the same queries to find out nothing did.
Grep Beats Every Benchmark
Marketing copy is cheap. So here's what independent benchmarks say.
Grep recently ranked #1 on every major deep research benchmark - DRACO, DeepSearchQA, and DeepResearch Bench - ahead of Perplexity, Google, OpenAI, Anthropic, and Nvidia. No other product has held #1 on more than one.
Same underlying frontier model as the rest of the field. 18.8 percentage points more accurate on DRACO. The gap is everything around the model - the routing, the experts, the memory, the execution.
For the full breakdown across all three benchmarks, see Grep is #1 on every major deep research benchmark.
What Else Shipped in V2
Beyond Brain itself, V2 brings a long list of changes across the product including:
- Inline report editing. Our biggest feature request - reports are now updated in-line with versioning when you ask follow-up questions instead of appending a new section.
- Richer widgets in reports. Charts, graphs, tables, profiles are all upgraded in Grep 2.0
- Work with any file. Drop in PDFs, spreadsheets, transcripts, or images and Brain reads them directly.
- Office Files. Turn your research into Powerpoint, Excel and Word docs.
- Dedicated Workspace. Every research job now has a dedicated workspace in the cloud with all the files created, whether that's images, models, data sets, ready for download.
Billing That Makes More Cents
With the launch of Grep 2.0, we're moving away from flat credit fees per low / medium / high effort request towards a usage based credit system than only charges you for the complexity of your task. A quick fact check uses a less credits. A full due diligence report uses more. A custom dashboard or 30-slide deck uses more than a one-page article. You only ever pay for what the work actually requires, based on the tokens Brain consumed to deliver it.
For a full breakdown of every shipped change, see the v2.0 Brain Launch entry in the changelog.
What Users Are Saying
Shopmonkey's underwriting team used to spend hours on each account review - pulling business records, checking litigation, scanning adverse media, verifying ownership across separate data sources. They tried Gemini to speed it up. It couldn't deliver the depth or structure they needed.
With Grep, the same research takes minutes. Every flagged account now runs through Brain before a credit decision.
When Stripe was about to shut down one of Shopmonkey's merchants for excessive disputes, Crystal Anderson, who leads Risk Strategy & Operations, ran the business through Grep. The report surfaced a $26K litigation record and a news clip that looked concerning, but the underlying context showed the merchant was legitimate. Crystal told Stripe to stand down.
There was no comparison. Grep does the summation, the visuals, the summaries automatically. It even flags insights. Grep was by far superior to Gemini.
Crystal Anderson, Risk Strategy & Operations, Shopmonkey
Read the full Shopmonkey case study.
Try Grep V2
V2 is rolling out now to all Grep users. Your workspace is already waiting.