01 / Support
How can we help?
Answers for every product in the suite, in one place. Pick a product below, or write to us directly.
Still stuck?
Write to us. Replies follow your plan's support targets: 48, 24 or 12 business hours.
Radar*
Research on the category, with a checkable source behind every claim.
With no key: GDELT for press tone and volume, publisher RSS feeds, SEC filings, open academic sources, Wikipedia and Wikidata, and the Meta Ad Library, which only returns ads delivered in the EU or UK. A paid tier adds entity-resolved news, audience data and creative examples once your organization sets a key for each provider.
ChatGPT on every plan, and Claude, Gemini and Perplexity on Agency and Enterprise or as a monthly add-on. They run on Signal*form's own accounts, so you add no API key. Google AI Overviews is not available yet. Every run asks the same unbranded questions a buyer would type, so the answers compare run to run.
Yes. Every section is written from sources and figures the model is given, and each claim points back to one of them. A citation that does not match a provided source is dropped from the report rather than kept with a caveat.
Not on the free tier. Reddit's free tier does not allow commercial use and X has no usable free tier, so neither is read. Social conversation and creative examples come in once your organization adds a paid provider.
Each report draws on your organization's monthly credits, which your plan includes. Before you run it, Radar* shows the cost in credits and what you have left, and offers a top-up if the balance is short.
Metric*
Every client's campaigns, down to the ad, in one report you can send.
Meta (Facebook and Instagram), TikTok, LinkedIn, YouTube, X (formerly Twitter), Google Ads and Pinterest connect through one account login. Google Analytics 4 and Shopify connect through their own login, each scoped to one client, and a member can map a Google Sheet or a BigQuery query column by column. Search Console, Klaviyo, Mailchimp and HubSpot are built and rolling out, so they aren't open to every account yet. Snapchat has no connection: its campaigns come in from a CSV upload, as any platform's export can.
Yes. Ad platforms, Google Analytics 4 and Shopify sync every night, and anyone on your team can press Sync on a connection for a fresher read.
Each platform's campaigns keep that platform's own reported numbers, side by side rather than blended into one figure. Attribution splits a connected store's orders across the campaigns clicked before them under the usual rules, and shows Google Analytics 4's own data-driven credit when the client's property is connected.
Yes. A client link, branded with the client's logo and color, shows only that client's campaigns. There is no username or password, and revoking the link ends access at once.
No. Every connection only reads from the platform; nothing in Metric* changes a budget, pauses a campaign or edits an ad. The AI analysis can recommend a budget change, but a person makes it.
Brain*
Ask about any client and get an answer with the exact source attached.
Not through anything the product offers. A client's portal reads only what you marked client-visible; vendors, their grades, private notes and the sharing log are never visible to a client. A test suite signs in as a client and checks that none of it is reachable, before every release.
Client content goes to Anthropic's API only when someone in your agency asks a question or requests an insight. Training and retention for those calls follow Anthropic's commercial API terms, not a setting in Brain*. Indexing documents for search sends their text to OpenAI's API for embedding only, under OpenAI's API terms.
PDF, Word, plain text, markdown, CSV and JSON or JSONL files uploaded per client; Google Drive files, imported once or from a folder Brain* watches and re-reads every hour; meetings recorded in Brain* or transcribed in Spark*; and Spark* briefs or Flow* projects someone brings in. There is no calendar or email connector in Brain*: everything comes from a file, a folder, a meeting or a record someone adds.
Yes, with one limit. Keyword search matches Spanish content at full strength and every record and citation displays correctly; ranking documents by meaning works better in English today.
Yes, and the portal tells the client so. Questions the portal could not answer come back to your team as findings, which is how you learn what a client wanted and could not find.
Spark*
A messy client call, turned into a brief the whole team can work from.
It helps the strategist. It removes the blank page and the formatting, and its review catches vague or unsupported claims, but the strategic judgment in the brief is still yours.
Yes. Record the call in the browser or upload the audio, and Spark* transcribes it and fills the intake from what was said. You check the fields before the draft is written.
Yes, as a house template. An admin defines it in Spark* › House templates, starting blank, from a built-in template or from an existing brief, with prefilled text and hints for the overview, objective, mandatories and success metrics, and can make one the default. It shapes one of the three brief types (creative, strategic and flash); a document or a custom layout can't be uploaded as a template.
Export the approved brief to PDF or share it on a client link, which opens without an account. Internal approval runs through the brief's own approval steps.
English and Spanish. Choose the language when you draft, review or generate a persona, and Spark* writes in it.
Mood*
One direction the whole team builds from, instead of a palette re-argued in every deck.
Image, video and instrumental-audio models from vendors including OpenAI, Google, ByteDance, xAI, Kuaishou and Ideogram. For each request, Mood* picks the model by the job, shows its cost and rights before you generate, and lets you switch any single generation.
It depends on the model, and Mood* tracks it per model. Every asset's status (cleared, restricted or unknown) is listed in the export's manifest, and exporting anything not cleared needs an explicit acknowledgement first.
Send two to four locked directions on a client link. The client opens it without an account, rates each direction, leaves a note or says "Not this one", and picks one; your team sees every answer under its direction.
They expire 30 days after they are made. Pinning a generation, or uploading a reference directly, keeps it.
No. Generating, pinning, the palette, locking a direction and the export all work on Mood* alone. Starting a board from a brief needs Spark*, and keeping a direction in a client record needs Brain*. On the Base plan, Mood* is an add-on; it is included in Pro and Agency.
Asset*
Every version of a file, the decision on it and what it did once it ran, in one place.
No. A review is shared on a client link; the client marks up, comments and records their decision there. You can revoke the link at any time.
It stacks onto the existing file rather than creating a second one. The history keeps every version with who uploaded it and its approval state, and the newest becomes current.
Numbered pins, rectangles and arrows on an image; a comment at a moment in a video; a note on a page of a document. Every mark stays on the version it was made on.
Through Metric*. Link a file to the ads it ran as; filename matches are suggested and every link is confirmed by a person. The file then shows the spend, cost per result and flight for those ads.
On images, in three places: Claude Sonnet 5.5 tags an image and reads the words on it, on upload or when you ask; Google's Nano Banana 2 resizes it into the channel sizes you pick, with the credits shown first; and search by meaning runs inside Signal*form. Comments, markup and decisions never go to an AI vendor, and every approval is a person's. On the Base plan, Asset* is an add-on; it is included in Pro and Agency.
Flow*
What the agency has committed to, who is doing it and whether it is making money.
Tasks, yes: a project imports an Asana, monday.com or ClickUp CSV export, or a plain sheet, with a preview of every row before anything is written. Clients, projects and the rate card have no import yet and are entered by hand; plan on a few hours for active work. Margin and the at-risk flag report real numbers once a project has a fee, a rate and a week of logged time.
Whoever your admin allows, by named role: Finance and Account Director see margins on every project, Project Manager sees delivery but not cost rates, and Contractor never sees financial data. The database enforces it, not just the menus.
Anyone who logs time or is assigned work needs a seat; the Contractor role limits it to their own bookings, tasks and time. A client does not need a seat: they open a client link with no account.
You can run projects and capacity without it, but not margin. Margin comes from logged time at a rate, so a project nobody logs time against shows no margin rather than a number to act on. Timesheet approval is off by default, so logging stays one step.
In sequence per organization, assigned when an invoice is approved, for example INV-2026-0042. An admin sets the prefix and the next number in Settings. Drafts have no number, and a number stays with its invoice once assigned.
Skill*
Who can do the work, and how well, before you have to ask around.
Everyone in your agency with access to Skill* sees colleagues' directory profiles, because that is what makes staffing work. Each person edits their own skills and history; only an admin sets someone's manager or level. Time off, reviews and pay are private to the person, their manager and whoever an admin allows.
Not today. Every page sits behind sign-in, and there is no public profile or share link.
Yes. One organization can define several agencies or offices alongside its departments, and profiles and services are tagged to one, so a group keeps one roster. Time off and onboarding are shared across the whole organization today.
No. It does not run payroll, tax or benefits. It replaces what agencies build around a thin HR directory: a spreadsheet of who speaks what, a policy wiki and staffing by memory. If you run an HRIS, Skill* sits beside it.
Yes. Profiles, the directory, the org chart, matching, onboarding, time off, reviews and recognition all work on Skill* alone. With Flow*, approved time off also reduces capacity there, and skills rank who Flow* proposes for a role.
Stack*
Every tool you pay for, who owns it and what it costs this month.
No. Every tool, cost and renewal date is entered by a person, in the app, through a CSV import or through the public request form, or read from a signed contract a person uploads and confirms. With Google Workspace connected, Stack* also reads who signs in to which apps to count seats in use; it never reads a charge. If a charge changes, someone updates it here too.
Import a CSV of your current tools, add them one at a time, or share the public request form so the people who use each tool can report it without an account.
Yes. Set how much of each tool serves each client and Cost by client shows every client's monthly share, across active tools only. Totals in different currencies are shown side by side, never added.
Every Monday a digest goes to the organization's owners and admins with the full list, and to each tool's owner with their own tools, grouped by what ends in the next 30, 60 and 90 days.
Everyone in your agency with access to Stack*; there is no separate finance-only view today. The public request form cannot read anything in the ledger.
Pulse*
Start the day knowing what is urgent, who is waiting on you and what you got done.
No. Every reply it drafts waits in Approvals, even if you tell the chat to send it. After you approve, you have 60 seconds to undo; then it shows as queued and sends within a minute.
Nobody. Your brief, tasks and drafts belong to your account alone, with no support access path to them. Your manager does not get a copy, and a colleague cannot search your mail.
Once a night: your last week of mail, replies you sent in the last two weeks, tomorrow's calendar and your open Google Tasks; with Flow*, your own tasks that are overdue or due soon. It does not read Slack or any chat app, never reads another person's account and never deletes, moves or archives mail.
No. There is no connection from Pulse* into Brain*, Flow* or any shared part of the suite, and the code refuses one. Nothing you read here reaches a shared view.
No. It connects to Google: Gmail, Calendar and Tasks. It doesn't connect to Outlook, Microsoft 365 or Teams.
Something not working? Write to us and a person replies within your plan's support target.