The technology is clearly solid, and the documentation is decent for covering the basics. The free trial was also easy to get started with. The Search API is the part I’d keep even at our scale, since the relevance on a per-query basis is strong. Review collected by and hosted on G2.com.

Everything about this product seems built for teams with dedicated data staff and a real budget. The pricing ramps up fast beyond what an internal-tools side project can reasonably justify, and a lot of the more interesting features are gated behind sales conversations. I also burned through the free pages in a single weekend of testing.

In practice, we use it for our production feed, where reliability really matters, and we rely on open source tools for more exploratory work. That split works well for us, but the marketing doesn’t make the enterprise focus clear upfront. Review collected by and hosted on G2.com.

We’re in financial services, so everything turns into a security review. Nimble got through ours faster than any data vendor we’ve evaluated. Their approach checked the boxes we care about: zero retention, PII masking, audit logs, and no training on our data. Their trust center also covered about 90% of our questionnaire before we even had to ask. What really made the security review manageable, though, was the search agents’ citations and the audit trail. Review collected by and hosted on G2.com.

The region-specific data residency options feel limited compared to some competitors. It worked for our use case, but if you have strict residency requirements, it’s worth double-checking whether the available regions meet your needs. Review collected by and hosted on G2.com.

I ran a build-vs.-buy analysis three separate times, trying to justify building an internal platform. Each time, Nimble came out ahead once I priced things honestly proxy costs, ongoing maintenance, and the constant anti-bot whack-a-mole. Their data quality also beat what we were getting from our internal scrapers. And our internal build never really solved discovery; their search agents consistently find sources we never would have thought to list. Review collected by and hosted on G2.com.

You’re trading some control for convenience. In certain edge cases with custom extraction, I’ve found you sometimes need help from their team instead of being able to resolve it with a configuration change on your side. Review collected by and hosted on G2.com.

The company research agent is better than I expected. I gave it a list of 50 target accounts and got back profiles that would have taken an intern weeks to put together. It cited sources for everything, which made verification quick and straightforward. Review collected by and hosted on G2.com.

5,000 trial pages sounds like a lot—until you point an agent at a real research task. We burned through the allowance in two days of evaluation and then had to ask for an extension. They did grant it, but I’d rather they start people off with more from the beginning. Review collected by and hosted on G2.com.

Feed it a company list and you get back enriched profiles pulled from live web data tech stack signals, hiring activity, news, and reviews. Our SDR team’s connect rate went up because the personalization is based on things that happened this week, not a stale database. What I like is that enrichment starts with a live web search on each account, rather than relying on a stale index. Review collected by and hosted on G2.com.

Matching companies to domains still requires a cleanup step on our side when names are ambiguous. If the API offered some fuzzy-matching support, it would save us from having to run a separate preprocessing job. Review collected by and hosted on G2.com.

Everyone talks about the agents, but the plain search endpoint is excellent. You get rich content in a single call, focus modes that actually matter, and pricing that feels sane. We’re using it as the retrieval layer for three internal tools now, and it’s been a solid fit for that role. Review collected by and hosted on G2.com.

My one real gripe is that Fast Mode is enterprise-only. We tried it during a sales evaluation and it was noticeably better for latency-sensitive work, but we can’t justify moving up to the enterprise tier just to get that one feature. Review collected by and hosted on G2.com.

The core extraction quality is high, and the Snowflake integration is well executed. When the managed feeds run, they run smoothly and deliver consistent results. Since the feeds are built on their search and retrieval capabilities under the hood, that relevance comes through clearly in the overall output quality. Review collected by and hosted on G2.com.

I had a rough billing experience early on. I was moved to a higher tier without clear notice, and there wasn’t a self-serve cancellation option, so I had to email and wait for a response. To their credit, they fixed it and credited the difference, and it hasn’t happened again. Still, the billing side of the product feels noticeably less polished than the data side. Also, there’s no support chat—it’s email only. Review collected by and hosted on G2.com.

Governance. Validation, schemas, audit logs, delivery SLAs, no training on our data. Every other vendor we evaluated felt like a tool you rent. Nimble feels closer to infrastructure, and our procurement and security teams picked up on that difference too. For ad hoc discovery the search agents give our teams relevant, cited sources. Review collected by and hosted on G2.com.

Honestly, there’s nothing major to complain about. If I had to pick something, I’d say the cookbook library could be larger, and the community around the product is still fairly small, so you can’t rely on Stack Overflow answers the way you can with Scrapy-based stacks. Review collected by and hosted on G2.com.

We benchmarked Nimble against Bright Data and Oxylabs across 50 protected retail targets. In our tests, Nimble delivered the highest first-attempt success rate and the lowest median latency, especially on JS-heavy pages. The VX8 render driver handled the anti-bot measures that stopped the others. Their live search using real browsers is also a major reason discovery on protected targets works at all. Review collected by and hosted on G2.com.

VX8 pricing is custom, and we had to get on a sales call just to receive a quote, which slowed down our evaluation. Their proxy pool also seems smaller than the big two, so when we needed very long-tail geos—specifically coverage in a few smaller LATAM countries—the available options were limited. Review collected by and hosted on G2.com.

When it works, the data is clean, and the success rates on hard targets are better than what we got with Decodo. The Generate Agent feature—where you describe what you want in plain English—is genuinely cool. What really helps is the search layer underneath, which is what makes the generated agents feel relevant rather than generic. Review collected by and hosted on G2.com.

It’s clearly built with larger deployments in mind. The $2500/month managed tier is more than we need, and the self-serve API takes a bit more engineering effort than the marketing suggests. The documentation also assumes you already know their vocabulary, which makes the learning curve steeper than it needs to be. We did look at cheaper options, but the data quality for our targets is what ultimately kept us here. Review collected by and hosted on G2.com.