Surfacer has assumed since launch that you'd work on one ICP at a time. Your discovery, your watch list, your blocklist, and the preferences the model picked up from your swipes all lived inside a single ideal-customer definition. This release lifts that limit. Workspaces let you keep several ICPs in the same account and switch between them from the sidebar. The companies page also gains a new search bar that finds anything in your library, anything other customers have already scraped, or kicks off a fresh scrape of a new company.
One workspace per ICP
Open the sidebar. Above the nav, a new dropdown shows the active workspace by name, colour, and icon. Click any other workspace to switch. The page reloads into that ICP context and the UI re-tints to its accent colour, so you always know which one you're in.
Each workspace owns:
- An ICP spec. Industry, region, size, free-text criteria, hard and soft rules.
- Discovery tasks. Every task you run is filed under the workspace that was active when you ran it.
- The monitored list. Your 100-slot watch list is per workspace, so a DACH SaaS ICP and an EU industrial ICP don't share an arbitrary slot pool.
- The blocklist. Passing on a company in one workspace doesn't hide it from another where it might actually fit.
- Learned preferences. The per-user preference model split into one row per workspace. The model that learned from your DACH SaaS swipes can't drift the EU industrial ranker.
Switching is instant
There's no save-and-reload step. When you click a workspace in the dropdown, the whole app rewires to the new context: the feed pulls cards scored against the new ICP, monitoring shows the new watch list, the dashboard's call list re-ranks, and the Live Signals stream filters down to the new monitored set. The nightly top-up cron also reads the active workspace, so switching late in the evening means the next morning's feed comes from the new ICP.
Per-workspace identity
When you create a workspace you pick a colour and an icon from a small set: green, amber, purple, and blue for colours; building, rocket, target, flask, factory, and a handful more for icons. The colour re-tints buttons, links, focus rings, and chart accents across the whole app whenever that workspace is active, and the icon appears in the sidebar avatar and the picker. They're small touches, but they pay back when you're juggling three ICPs in an afternoon and need to recognise the active one at a glance.
Quick search
The Companies page has a new search field anchored to the top right of the header. Type two characters and it queries three places at once.
- Your library. Anything you've already saved or rated. Click to open the detail panel inline.
- The global pool. Companies other Surfacer customers have already scraped, indexed by name and domain. Click Save and it joins your library instantly. No fresh crawl, no LLM cost.
- Scrape on demand. If nothing matches and your input looks like a URL, you can scrape it. If your input is a plain name, we Google-search the URL first, then scrape. One credit per scrape, regardless of which entry point.
Inherit-on-scrape
The shared-pool branch is quietly the most useful of the three. Surfacer has spent the last year building a shared index of every company any customer has ever scraped, which now holds tens of thousands of records. When your search hits something already in that index, saving it to your library is free. There's no re-scrape, no verifier, and no waiting; you inherit the full record (identity, address, size, jobs, products, news, signals, growth) and we link it into the active workspace.
This is the same inherit-the-cache mechanic that makes the atoms composer in 0.4 cheap to run for newly-joined customers. Every customer adds to the pool, and every customer benefits from it.
Why both ship together
The two features are deliberate complements. A multi-workspace account quickly raises the question of how to find a particular company across several workspaces, and a quick-search bar without workspaces would leave heavier users with the older problem of a single account turning into a graveyard of conflicting ICPs. Shipping both at the same time closes both gaps: users juggling several books of business get the multi-workspace model, and everyone else, regardless of how many workspaces they keep, gets one input that finds any company in the system.
Also in this release
- New S2 mark. The favicon and wordmark refresh to a cleaner squircle silhouette.
- HubSpot blocklist sync per workspace. Existing-customer and competitor imports now scope to the workspace active at import time.
- Settings reshuffle. The ICP tab is now Current ICP, and a new Workspaces tab gives you the full editor for renaming, recolouring, and archiving.
What's next
Quality remains the metric we're focused on, and the shape of the next two releases is about depth. The first push is into deeper signal integrations. Most of what we surface today is computed from the public web, and the next wave brings first-party connections to the systems where buying intent actually shows up: job-board APIs we don't have to scrape, pricing and catalogue feeds, patent filings, and funding-round and leadership-change registries pulled straight from the registry rather than inferred from press. Each of those tightens the precision band on a signal family we already ship.
The second push is observability over time. A monitored company isn't a row, it's a timeline. The next dashboard surfaces a company's last twelve months of signals as a single continuous strip, so the pattern is legible at a glance: whether hiring is accelerating or rolling back, whether pricing changes are isolated edits or part of a coordinated rebrand, whether news mentions cluster around a single quarter or spread evenly across the year. The same data we already collect, finally rendered in a way that answers what's changing rather than what changed last night.