Claude Bronto connectorWhat Claude can do in your Bronto account.
The Claude Bronto connector exposes 16 tools: 14 read, 2 write. Claude can explore your datasets, search and chart logs, check monitors and coverage, then save a search or create a muted monitor. Below: what it reads, what it creates, and the admin switch.
Verified Trustpilot reviews · AI, automation & growth agency
Why plug Bronto into Claude
During an incident, the slow part is rarely the search itself: it is knowing which dataset holds the answer, which field to filter on and how to write the query. With the connector, you ask what is happening in production in plain English, and Claude walks through your Bronto data the way an engineer would, from dataset discovery to a precise search.
From a vague symptom to the failing route. Users say the site feels slow. Claude uses get_datasets to find the CDN logs, get_keys and get_key_values to learn the fields, then timeseries to compare response times hour against hour and point at the worst paths.
An alert that is ready next time. Once a query has exposed the problem, create_monitor turns it into a monitor, with an investigation prompt Bronto runs when it fires. It is always created muted, so nobody gets paged before you review it.
Blind spots, listed. get_coverage reports which datasets are ingesting but unmonitored or never searched, and validate_monitors re-runs each monitor's query to catch alarms that broke silently.
What it won't do: edit or delete an existing monitor, unmute one, or change ingestion. No tool on the sheet does that. Nothing runs by itself either: Claude searches when you ask, the monitors you create are what watch in between. Wiring alerts into other apps is an automation tool's trade; the Integrations hub covers that side.
The vocabulary in one minute
Five words you will meet while plugging Bronto into Claude.
- Connector
- The link you set up once between Claude and an account you already have, so Claude can query it while it answers you.
- Tool
- One named job the connector opens to Claude, such as searching logs. Claude chains the ones it needs; the directory lists them by name.
- Authorization
- Bronto's own sign-in screen, where you grant Claude the access and scopes it will use afterwards. Given once per person, revocable.
- Approval
- The confirmation Claude waits for before it goes ahead with something that changes your account, shown in the chat when it matters.
- MCP
- The shared standard connectors are built on: it is what lets an assistant like Claude talk to an outside service such as Bronto.
Connect Bronto to Claude in three steps
- 01
Find Bronto in Claude
In Claude's settings, open Customize, then Connectors, and look for Bronto. On a Team or Enterprise workspace, an Owner or Primary Owner must switch the connector on before each engineer can link it.
- 02
Start the connection
Click Connect on its row, then sign in to Bronto in the window Bronto opens, with a login method your admin allows. If the link breaks later, Disconnect and connect again.
- 03
Review the scopes
Read the authorization screen before you approve. It is Bronto's screen, not Claude's, and it lists the permission scopes the access covers. You can withdraw that access later from Bronto's side too.
The 16 tools, sorted by what they do
Bronto gives Claude 16 tools: 14 that read your account, 2 that change something in it.
Fourteen tools read your data, two create something in your account. Names stay in English, exactly as Claude shows them.
- 14 read
- 2 write
What Claude reads (14)
14 toolsFourteen tools that discover, search, aggregate and audit without changing anything.
get_all_datasets_keys
Returns the field names of every dataset you can access in one pass, so Claude sees the schema across the board before it builds any search or aggregation.
get_coverage
Reports, dataset by dataset, whether it is monitored, searched and still ingesting. The result is a list of blind spots, such as a live dataset with no alert watching it.
get_datasets
Lists every available dataset with its name, collection, unique ID and tags. It is usually the first step of an investigation, when the right source isn't obvious yet.
get_datasets_by_name
Finds a dataset from its exact name and collection name, for when you already know what you are looking for and want Claude on that one source.
get_error_summary
Gives org-wide error and warning totals with a breakdown per dataset. Bronto precomputes it, so it costs no search quota.
get_key_values
Returns the values a given field takes in a given dataset, which lets Claude build exact filters instead of approximate ones.
get_keys
Lists the field names of one specific dataset, so Claude can inspect its schema before it queries that source and avoid filtering on a field that doesn't exist.
get_saved_searches
Lists and reads your team's saved searches, filtered by text, tag or field. Claude checks what already exists before proposing to add another one.
get_usage
Reports ingestion and search usage per dataset against your monthly limits, and flags whether each dataset is actively ingesting or dormant.
run_saved_search
Executes a saved search by its ID and returns the same results as running the query directly, optionally over another time window.
search_logs
Searches raw log events in one or more datasets over a time range, with filters, limits and sorting. It is the event-level tool for troubleshooting.
search_monitors
Finds existing monitors and, for a given one, its recent trigger and recovery history. Useful for a retrospective and to avoid duplicate alerts.
timeseries
Computes aggregated metrics and trends from log data, with counts, sums, averages, percentiles, grouping, time buckets and comparison windows.
validate_monitors
Re-runs the backing query of each existing monitor to spot alarms that broke silently or stopped receiving data.
What Claude creates (2)
2 toolsTwo tools that add a saved search or a monitor to your account. No source details a confirmation for them: the general rule below applies.
create_monitor
Approval: see the ruleAdds a monitor to your account, with threshold or no-data alerting and an optional investigation prompt Bronto runs when it fires.
create_saved_search
Approval: see the ruleSaves one or several queries for the team, with how each one renders: list, time series, table, top-list, pie, treemap or a geomap by country.
What Claude asks before it acts
By default, Claude stops and asks for your go-ahead before any action it takes on an account on your behalf. No source says whether creating a monitor or a saved search triggers a confirmation of its own.
Bronto adds a safety net of its own: every monitor Claude creates is muted, so nothing pages anyone until a person reviews it in Bronto. On the Bronto side, admins decide whether MCP access is allowed at all, which login methods count and therefore which users can connect. On Team and Enterprise, Claude's owners can let members skip some prompts, or restrict what a connector may do for the whole organization. Claude works with the connected person's rights, nothing more.
Which plans it's available on
Of the 819 sheets in the official directory, none shows plan availability. That answer is not published for any single connector: a genuine gap in the catalogue.
What is published is the general rule: remote connectors are open to all users on Claude, Cowork, Claude Desktop and mobile. On Team and Enterprise, an Owner or Primary Owner switches the connector on for the organization before members can link it. Bronto, for its part, documents hosted MCP access as open to all Bronto accounts. For this connector's current state in Claude, its sheet in the official directory is the reference.
Where this connector stops
A connector is not an automation. Claude calls these tools while it answers you: no search runs on its own at night, only the monitors you create keep watch.
Everything Bronto-specific here comes from Bronto's documentation: no Claude help article covers this connector. Two tools write, but the sheet stops at creation; changing or deleting a monitor is not among them. The directory badge is not a security audit, and Anthropic says on every sheet that it doesn't control which tools a publisher exposes or guarantee their behavior. For a comparison with another connector, the Claude adobe-workfront connector page follows the same structure.
Need help connecting Bronto to Claude?
A person reads every message.
