Front Search Explained (2026): Operators, Filters & the in:comments Gotcha

How search works in Front

Open search by clicking the search icon at the top of your Front inbox. As Front's guide to searching in Front explains, you can type words directly into the search pop-up to search across all parameters for a matching term, or narrow things with filters. Type your keywords and Front shows automatic suggestions in a dropdown; press Enter to see the full result list.

There are three ways to build a query, and they layer on top of each other:

1. Plain keywords. Type refund and Front matches the subject and body of conversations. Multiple words are treated as AND logic — refund urgent returns conversations containing both.
2. Operators. Type a colon-prefixed operator like from:, tag:, or is:open to filter precisely. Front prompts you to "Type + for advanced search" to surface the operator menu.
3. The search builder. Click the filter icon in the search pop-up to open a point-and-click builder that exposes every filter without you having to remember the syntax. After any search, a filter bar also appears so you can refine results further with a click.

The operators and the builder are the same thing wearing two faces — the builder just writes the operator syntax for you.

The operator cheat sheet

Front's operators fall into four families. The Front API search reference and the help center between them document the full set; here are the ones that earn their keep.

Message content — where the text lives:

• subject:invoice — match the subject line only
• body:invoice — match the message body only
• in:messages — text inside message bodies
• in:attachments — text inside attached PDF, CSV, XLS(X), DOCX, or TXT files (emails only); not-in:attachments excludes them
• in:comments — text inside internal comments (more on this below — it's the big gotcha)

People — who is involved:

• from:[email protected] — sender only
• to:, cc:, bcc: — the respective recipient fields; recipient: matches any of them
• with:acme — a contact involved anywhere in the conversation
• account:Acme Corp — contacts linked to a specific account

Teammates — who on your side touched it:

• assignee:priya — assigned to a teammate
• author:priya, commenter:priya, mention:priya, participant:priya — authored, commented on, was @-mentioned in, or otherwise participated

Properties — state and metadata:

• is:open, is:archived, is:snoozed, is:assigned, is:unassigned, is:unread, is:unreplied — conversation status
• is:spam, is:trashed — the two scopes everyone forgets (below)
• tag:vip — a tag; multi-word tags use underscores, e.g. tag:everything_is_awesome
• has:attachment, has:reminder, has:no-tags — presence filters
• before:2026-06-01, after:2026-05-01, during:2026-05 — date scoping (YYYY-MM-DD, or YYYY-MM for a month)

You stack operators for AND logic: tag:vip is:open assignee:priya returns open, VIP-tagged conversations assigned to Priya. That composability is where Front search stops being a text box and starts being a query language.

The in:comments gotcha (and trash and spam)

Here is the behaviour that trips up almost every team at least once. A plain keyword search does not look inside internal comments. When you type a word into the search pop-up, Front matches messages — the emails and replies — but the private @-mention discussion your team had about the conversation is not in scope by default. If a colleague wrote "escalate this to billing, the customer was double-charged" as a comment and you search double-charged, you may get nothing.

The fix is the in:comments operator. Add it — in:comments double-charged — and Front searches the comment text specifically. This matters enormously on a collaborative shared inbox where a lot of the real institutional knowledge lives in comments and mentions, not in the customer-facing thread. If you've ever been sure a decision was recorded somewhere in Front and couldn't find it, this operator is usually why.

The second silent default is the same shape. Per Front's help center, "The Trash and Spam folders are not automatically included when you search for keywords." A conversation someone trashed, or that Front flagged as spam, simply won't appear in normal results. To reach it you scope it back in with is:trashed or is:spam. The classic case: a customer swears they emailed you, the message got auto-filed as spam, and your agent's search comes up empty — because spam was never in the search to begin with.

Exact phrases, exclusions, and wildcards

Beyond operators, Front supports the refinements you'd expect from a serious search box:

• Exact phrases go in quotation marks: "weekly team update" matches that phrase in order, not the three words scattered apart.
• Exclusions use a leading dash: meeting -budget finds conversations about "meeting" but not "budget." You can exclude senders (-from:[email protected]) and tags too.
• Prefix wildcards use an asterisk: soph matches "Sophia" or "Sophie." One honest caveat — prefix search does not combine with field modifiers, so body:soph won't behave as a wildcard.

Saved searches

Any query you'll run again is worth saving. Once you've built a search — say is:open is:unreplied tag:vip for "VIP conversations still waiting on us" — Front lets you save it so it becomes a one-click view instead of a query you retype. Saved searches are how support teams keep a live triage lens ("unassigned and older than a day"), how sales keeps "open deals I'm a participant on," and how ops keeps a compliance filter permanently at hand. They turn search from a lookup into a dashboard.

The honest limits — and where an AI layer picks up

Front's search is genuinely strong: fast, operator-rich, and precise once you know the syntax. For finding things, it's hard to fault. But notice what it is and isn't.

Search matches strings, not meaning. refund finds the word "refund"; it doesn't understand that "I was double-charged and want my money back" is the same intent expressed without the keyword. It can locate the conversation, but it can't read it, reason about it, or answer it — search is a retrieval tool, not a resolution tool. And a few adjacent capabilities are plan- or add-on-gated rather than part of search itself: Front's richer Analytics reporting scales up with tier (formal SLA-attainment dashboards land on the larger/enterprise plans), and Front's Autopilot agent is a paid add-on rather than something included in the base plan. Front's pricing tiers lay out exactly what unlocks where, and Front analytics explained covers the reporting side.

This is the seam where an AI agent layer fits — not to replace search, but to do the reasoning-heavy work retrieval structurally can't. The category of AI agents for customer service exists precisely for the part a string match can't do. Macha is one such layer: it runs on top of the Front you already use through the live Macha–Front connector — it does not replace Front, your inboxes, your rules, or your search. You keep search doing what it's good at: pinpointing conversations by sender, tag, date, or phrase. Then Macha's agent reads the conversation it found, understands intent rather than keywords, and drafts or sends a grounded reply — pulling a real order or account status through a custom tool that turns your REST API into something the agent can call. Search finds the double-charge thread; an agent then actually resolves it. If you want the mechanics, connecting Front to Macha to route conversations to AI walks through it, and Macha's credits are consumed per AI action, never per resolution — retrieval and reasoning have different costs, and it's honest to price them that way.

The clean division of labour: let Front search be the precision retrieval layer, and layer an agent on top for the part a keyword match can't do — understanding the conversation and answering it well.

FAQ

Does Front search my internal comments? Not by default. A plain keyword search matches messages (emails and replies) but skips private comments. To search comment text, add the in:comments operator — for example in:comments escalation. This is the single most common reason a team thinks something "isn't in Front" when it is.

Why don't trashed or spam conversations show up in search? Front deliberately excludes the Trash and Spam folders from keyword searches. Add is:trashed or is:spam to your query to include them. This is why a customer's email that got auto-filed as spam can look "missing" until you scope spam back in.

How do I search for an exact phrase in Front? Wrap it in quotation marks: "order confirmation" matches that exact phrase rather than the words individually. Combine it with operators — "order confirmation" is:open — for tighter results.

Can I save a search in Front? Yes. Build the query you want (operators included), then save it so it becomes a reusable one-click view. Saved searches work well as live triage lenses like "open and unreplied" or "VIP still waiting."

Can I add AI to Front search without replacing Front? Yes. An AI agent layer like Macha connects to Front as a live connector and runs on top of your existing inboxes and search — it doesn't replace them. Search keeps finding conversations; the agent reads the found conversation, understands intent, and drafts or sends a grounded reply.

Ready to turn "found the conversation" into "answered it"? Start a free trial of Macha and connect it to your Front in minutes.