> ## Documentation Index
> Fetch the complete documentation index at: https://docs.os.default.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Default MCP server

> Connect Claude, Cursor, Codex, VS Code, or another MCP-compatible AI client to your Default workspace with the Default MCP server (beta): setup, tools, access, and troubleshooting.

The Default MCP server connects an MCP-compatible AI client to approved data and tools in your
Default workspace. MCP, the Model Context Protocol, is a standard way for AI clients to call
external tools.

Use it when you want an external assistant or custom agent to ask questions about your
workspace. It goes through the same permissioned access layer that Default uses for agent
workflows.

> [!NOTE]
>
> **Admins and owners can use the enabled MCP tools.** Members and editors can also use the
> booking-link tools.
>
> Booking tools use the member's existing event and queue access. They do not grant access to
> more events or queues. Routing changes need a separate MCP routing-write grant and the same
> queue or policy permissions as the app. Viewers cannot use booking tools.

## Get started

Connect your client to the Default MCP server:

```text Server URL theme={null}
https://mcp.default.com/mcp
```

Every client uses the same 2 steps: add the server URL, then sign in to Default through the
browser-based OAuth prompt with your Default account. There is no access-token, bearer-token, or
API-key method.

> [tab] Cursor
>
> [![Install the Default MCP server in Cursor](https://cursor.com/deeplink/mcp-install-dark.svg)](cursor://anysphere.cursor-deeplink/mcp/install?name=default\&config=eyJ1cmwiOiJodHRwczovL21jcC5kZWZhdWx0LmNvbS9tY3AifQ%3D%3D)
>
> Select **Install MCP Server** to open Cursor and add the Default MCP server, then complete
> the sign-in prompt. To configure it manually instead, add this to `~/.cursor/mcp.json`:
>
> ```json ~/.cursor/mcp.json theme={null}
> {
> 	"mcpServers": {
> 		"default": {
> 			"url": "https://mcp.default.com/mcp"
> 		}
> 	}
> }
> ```

> [tab] Claude Code
>
> Run this command to add the Default MCP server:
>
> ```bash theme={null}
> claude mcp add --transport http default https://mcp.default.com/mcp
> ```
>
> Then, inside Claude Code, run `/mcp` and follow the sign-in prompt.

> [tab] Claude
>
> In Claude on the web or in the Claude desktop app:
>
> 1. Open **Settings**, then **Connectors**.
> 2. Select **Add custom connector**.
> 3. Enter a name, such as `Default`, and the server URL: `https://mcp.default.com/mcp`.
> 4. Select **Add**, then complete the sign-in prompt.
>
> Custom connectors require a paid Claude plan.

> [tab] Codex
>
> Add the Default MCP server to `~/.codex/config.toml`:
>
> ```toml ~/.codex/config.toml theme={null}
> [mcp_servers.default]
> url = "https://mcp.default.com/mcp"
> ```
>
> Then sign in:
>
> ```bash theme={null}
> codex mcp login default
> ```

> [tab] VS Code
>
> [![Install the Default MCP server in VS Code](https://img.shields.io/badge/VS_Code-Install_Default_MCP-0098FF?style=for-the-badge\&logo=githubcopilot\&logoColor=white)](https://vscode.dev/redirect/mcp/install?name=default\&config=%7B%22type%22%3A%22http%22%2C%22url%22%3A%22https%3A%2F%2Fround-lake.dustinice.workers.dev%3A443%2Fhttps%2Fmcp.default.com%2Fmcp%22%7D)
>
> Select the button to open VS Code and add the Default MCP server, then complete the sign-in
> prompt. To configure it manually instead, add this to `.vscode/mcp.json` in your workspace:
>
> ```json .vscode/mcp.json theme={null}
> {
> 	"servers": {
> 		"default": {
> 			"type": "http",
> 			"url": "https://mcp.default.com/mcp"
> 		}
> 	}
> }
> ```

> [tab] Other clients
>
> MCP is an open protocol with many compatible clients. In any client that supports remote MCP
> servers over Streamable HTTP:
>
> 1. Add a new remote MCP server with the URL `https://mcp.default.com/mcp`.
> 2. Complete the browser-based OAuth sign-in with your Default account.
> 3. Refresh the tool list.
>
> Many clients accept this common configuration shape:
>
> ```json theme={null}
> {
> 	"mcpServers": {
> 		"default": {
> 			"url": "https://mcp.default.com/mcp"
> 		}
> 	}
> }
> ```

After you connect, ask your client to list or refresh MCP tools, then start with a specific
question: "What company fields can I query?" or "How many companies match this filter?" Your
role, connected integrations, and enabled workspace features filter the tool list.

## What you can do

Common use cases include:

* Inspect your company and person schema before you ask a data question.
* Count or query company and person records in your Default data model.
* Count or query Salesforce and HubSpot records through your connected CRM.
* Resolve a teammate to their CRM user ID so you can filter records by owner.
* Update a single CRM record through a confirm-before-write flow.
* Start an enrichment run for records in a table view or workflow.
* Browse booked meetings, event types, and each rep's availability setup. Check open slots and
  create or revoke booking links.
* Explain why a lead routed to a rep, and manage queue members and routing policies.
* Build, check, and publish workflows, with a preview you approve before each save. See
  [Build workflows with an AI agent](https://docs.os.default.com/mcp-workflow-building.md).
* Inspect workflow runs and read the logs for one run.
* Check integration health and find reps missing a CRM mapping or a calendar.
* Upload a CSV, map its columns to Default fields, and query it as a segment.
* Give external AI workflows a standard way to talk to Default without copying data out of
  your workspace.

The exact tools you see depend on your role, grants, connected integrations, and enabled features.

## Docs MCP server

Default also runs a second, separate MCP server for this documentation site:
`https://docs.os.default.com/mcp`.

It searches and reads the published docs, so your assistant can answer "how do I" questions
with current product documentation instead of stale training data. It is public: it needs no
sign-in and no workspace role, and it never reads or changes your workspace data. Its only
write tool sends feedback to the docs team.

| Tool | Type | What it does |
| - | - | - |
| `search_default` | Read | Searches the documentation and returns the relevant pages. |
| `query_docs_filesystem_default` | Read | Browses the documentation as a read-only file tree. |
| `submit_feedback` | Write | Reports an incorrect or confusing page to the Default docs team. |

> [!TIP]
>
> Add both servers to your client. The docs server answers questions about how Default works.
> The workspace server, described on the rest of this page, answers questions about your data.

## Available tools

Most tools below require the admin or owner role. Members and editors can use
`events_list`, `event_get`, `host_options_list`, `scheduling_links_list`,
`scheduling_link_create`, `scheduling_link_revoke`, `scheduling_link_update`, `my_booking_link`, `suggest_times`, `meetings_list`, `meeting_get`, `meeting_cancel`, `meeting_mark_no_show`, `meeting_reschedule`, `meeting_reassign`, `meeting_add_guests`, and `meeting_mark_held`.

Routing write tools require a separate MCP routing-write grant and app permissions.
Other tool families keep their existing restrictions.

Read tools return data. Write and run tools change data or start jobs, so treat them accordingly.

> [!NOTE]
>
> The MCP server is in **beta**. The tools below are on for every workspace with MCP access,
> except the three noted under [More tool families](#more-tool-families): enrichment, data import
> and segments, and `run_workflow`. Those turn on per workspace.

### System

| Tool | Type | What it does |
| - | - | - |
| `system_ping` | Read | Confirms the MCP service is reachable and that its authenticated link to Default's API works. |
| `calculate` | Read | Evaluates an arithmetic expression and returns the exact result, such as percentages, ratios, and sums. |
| `submit_feedback` | Write | Sends feedback about this MCP server to the Default team: a bug, a missing capability, a wrong result, or praise. |

> [accordion] calculate details
>
> Supports `+ - * / % ^`, parentheses, unary minus, and the functions `abs`, `round`, `floor`,
> `ceil`, `sqrt`, `min`, `max`, and `pow`. It accepts numbers and operators only, not variables.
>
> Example: `(150000 * 0.85) / 12`.

### Company & person data

These tools read your Default data model. The data model holds the company and person records
that fill your Tables.

| Tool | Type | What it does |
| - | - | - |
| `data_schema` | Read | Returns the field descriptors for `person` or `company` records so you know what you can filter on. |
| `data_count` | Read | Returns an aggregate count of `person` or `company` records, with optional schema-backed filters. |
| `data_query` | Read | Returns `person` or `company` rows with required pagination and a schema-backed field allowlist. |

> [!TIP]
>
> Call `data_schema` first so your client knows which fields exist before it filters or sorts with
> `data_count` or `data_query`.

### CRM (Salesforce & HubSpot)

These tools read and write through your connected CRM integration. Connect and sync the
relevant integration, Salesforce or HubSpot, first.

| Tool | Type | What it does |
| - | - | - |
| `crm_schema` | Read | Returns provider-aware field descriptors for a Salesforce or HubSpot object. |
| `crm_count` | Read | Returns an aggregate count for a CRM object, with optional schema-backed filters. |
| `crm_query` | Read | Returns CRM rows with required pagination and a schema-backed field allowlist. |
| `crm_actor` | Read | Returns the signed-in user's own Salesforce / HubSpot user IDs, for owner-filtered queries. |
| `integration_user_mappings_list` | Read | Lists CRM user mappings for the org. Filter by email or name to resolve a teammate before an owner filter. |
| `crm_write` | Write | Updates fields on a single CRM record in 2 steps: preview first, then re-call with `confirm: true` to commit. |

> [accordion] How crm\_write confirmation works
>
> `crm_write` never changes data on the first call.
>
> 1. Call it without `confirm`, or with `confirm: false`, to get a `pending_confirmation`
>    preview of the resolved change.
> 2. Show that preview to the operator.
> 3. After the operator approves, re-call with `confirm: true`.
>
> Default validates field ids against `crm_schema`.

## More tool families

The families below are on for every workspace with MCP access, with three exceptions that turn
on **per workspace**: enrichment, data import and segments, and `run_workflow` under Workflows.
If one of those is not enabled for your workspace, its tools do not appear in your client's
tool list. Ask your Default contact to enable it.

The routing write tools and the booking link tools use the same preview-then-confirm pattern
as `crm_write`: the first call returns a preview and changes nothing, and only a re-call with
`confirm: true` commits.
The data import tools use their own review step instead, described in that section.

### Enrichment

| Tool | Type | What it does |
| - | - | - |
| `enrichment_options` | Read | Discover provider IDs, supported fields, and organization waterfall IDs for `enrichment_run`. |
| `enrichment_run` | Run | Starts an enrichment job for records in a table view or workflow, using a provider or waterfall. |

> [accordion] enrichment\_run details
>
> Call `enrichment_options` with `object_type` first. Pass `providers[].provider_id` as
> `provider_id`, or `waterfalls[].waterfall_id` as `waterfall_id`. Use `nextCursor` to
> page through waterfalls in stable ID order. Both tools require the **Admin** role, and
> Default turns them on per workspace. The providers you can use are the ones enabled for your
> workspace, whether or not they are in a waterfall.
> Waterfalls retain their saved provider order.
>
> You have 2 ways to select records:
>
> * Pass explicit `record_ids`.
> * Omit them and set a `mode`, for example `all` or `all_empty_cells`, to resolve the record
>   set from the view.
>
> For AI-research providers, also pass a `prompt` or an `output_schema` that describes the
> research ask.
>
> This starts a job and spends enrichment credits. It does not return enriched values inline.

### Meetings & scheduling

Tools over your booked meetings, event types, booking links, open slots, and each rep's
availability setup.

| Tool | Type | What it does |
| - | - | - |
| `meetings_list` | Read | Browses booked meetings with hosts, attendees, the event, and the queue that booked each one. |
| `meeting_get` | Read | Reads one meeting: full host set, attendees, conferencing type, and cancellation reason. |
| `meeting_cancel` | Write | Previews and cancels a meeting the person can manage. Guests get the normal cancellation email. |
| `meeting_mark_no_show` | Write | Previews and marks a meeting as a no-show. The preview names the hosts who get their queue spot back. |
| `meeting_reschedule` | Write | Previews and moves a meeting to a new start. Guests get the updated calendar invite. |
| `meeting_reassign` | Write | Previews and hands a meeting to a new host: a person, or a queue that picks one at confirmation. |
| `meeting_add_guests` | Write | Previews and adds guests to a meeting. Each new guest gets the calendar invite. |
| `meeting_mark_held` | Write | Previews and marks a started meeting as held. The calendar event stays and no message is sent. |
| `events_list` | Read | Lists bookable event types with duration, conferencing type, and active state. |
| `event_get` | Read | Reads one event type's booking config, host lineup, booking-form questions, and linked workflows. |
| `event_update` | Write | Edits an event type's configuration, host lineup, and booking form in 2 steps: preview first, then re-call with `confirm: true`. Hosts and questions are full replacements. |
| `host_options_list` | Read | Lists members who can host meetings, with calendar health, and the queues you can use as hosts. |
| `scheduling_links_list` | Read | Lists your booking links across accessible events, or one event’s links; supports host, creator, state, and kind filters, booking URLs, compact/full detail, and pagination. |
| `available_slots` | Run | Returns the open booking slots for an event type or one booking link, over a date window. |
| `suggest_times` | Read | Finds open times for an event type or booking link without opening a booking session. Use it to suggest meeting times in a message. |
| `scheduling_link_create` | Write | Previews and creates a custom link with an expiry or fallback host; use my\_booking\_link for lasting links. |
| `scheduling_link_revoke` | Write | Previews and revokes a link; delete: true also frees its identifier. Admins and owners may pass up to 50 linkIds for an atomic batch. |
| `scheduling_link_update` | Write | Changes a link’s label, expiry, or visibility immediately; host changes require confirmation. |
| `scheduling_links_audit` | Read | Admins and owners: finds duplicate links, departed hosts, links never booked, and old private links. Returns counts and up to 20 compact rows per check; extraCount includes duplicate extras omitted by the limit. |
| `my_booking_link` | Write | Returns an existing booking URL for named hosts or creates a lasting link. Defaults to you. |
| `availability_get` | Read | Reads a rep's configured buffers, notice period, working hours, and out-of-office blocks. |
| `work_schedule_get` | Read | Reads a rep's working hours and the timezone they are expressed in. |
| `ooo_list` | Read | Lists a rep's out-of-office blocks, optionally within a date window. |

Admin example: “Find duplicate booking links on the 'Sales intro' event and revoke the extras.” The audit compares host identities and flags, case-sensitive trimmed names, visibility, identifiers, enabled and disabled saved URL parameters, and expiry days in the requested `timeZone`. Pass your browser’s IANA timezone to match the app’s local-day comparison; when omitted, the server’s local timezone is used and returned in the result. Links with fallback hosts are excluded from duplicate groups. The keeper is the oldest link with a non-canceled booking, or the oldest link if none has bookings.

Example: “Show me my booking links.”

> [accordion] available\_slots and booking link details
>
> `available_slots` is not a free read. Each call opens a booking session, and when a host is a
> queue that assigns one rep per booking, that opens a routing assignment you can see in
> `list_routing_activity`. Ask once over the range you care about instead of polling.
>
> To suggest times in a message, use `suggest_times`. It reads open times and opens no booking
> session and no routing assignment.
>
> An empty result means no open slot in the window, not a broken event. Check
> `availability_get` for the hosts' working hours and out-of-office before you report a problem.
>
> `scheduling_links_list` includes both kinds by default and returns compact rows with URLs; use `detail: "full"` for all fields. `scheduling_link_create` takes person and queue host IDs from `host_options_list`. A link with no expiry set expires after 7 days. `scheduling_link_revoke` is
> permanent: there is no un-revoke, and meetings already booked through the link keep their
> time.

### Queues & routing

Read tools that answer "why did this lead go to this rep" and "who is up next".

| Tool | Type | What it does |
| - | - | - |
| `list_queues` | Read | Lists routing queues with active state, policy, distribution rule, and assignment totals. |
| `get_queue` | Read | Reads one queue's members with weights, paused state, and the live up-next ranking. |
| `list_routing_policies` | Read | Lists routing policies with distribution rule, tiebreaker, and reset frequency. |
| `get_routing_policy` | Read | Reads one policy's active configuration and the queues that use it. |
| `explain_assignment` | Read | Returns the full decision record for one assignment: winner, scores, and excluded members. |
| `lookup_assignments_by_email` | Read | Finds a lead's routing history by email, including failed routing requests. |
| `list_assignments` | Read | Browses the assignment log: which lead went to which rep in which queue, and when. |
| `list_routing_activity` | Read | Browses the routing configuration changelog: who changed what, when, and why. |

### Routing writes

Write tools that change live routing behavior. They roll out separately from the routing read
tools, and every one of them previews first and commits only on `confirm: true`.

| Tool | Type | What it does |
| - | - | - |
| `queue_member_update` | Write | Pauses or unpauses a queue member, or changes their routing weight. |
| `queue_member_calibrate` | Write | Credits or debits the effective assignment count of one or more members of a queue, with a required reason. Several members share one preview and one approval. |
| `queue_member_add` | Write | Adds members to a routing queue. |
| `queue_member_remove` | Write | Removes members from a routing queue. Their historical assignments stay. |
| `routing_policy_update` | Write | Updates a policy's rules. The change applies to every queue using the policy. |
| `routing_policy_reset` | Write | Resets a policy's assignment counters to zero across all of its queues. |

### Rules of Engagement

A ruleset decides where a lead goes (a queue, a person, or no one) from its input values. These
tools read a ruleset in plain words, test it with sample leads, and change it.

| Tool | Type | What it does |
| - | - | - |
| `roe_rulesets_list` | Read | Lists rulesets with the live version number, rule count, and whether the draft has unpublished changes. |
| `roe_ruleset_get` | Read | Reads one ruleset: its rules in order, destinations by queue and person name, and the workflows that use it. |
| `roe_evaluations_list` | Read | Reads the evaluation log: which rule matched for a real lead, and where the lead went. |
| `roe_ruleset_test` | Read | Runs sample leads through the live version or the draft and returns the matched rule and destination. Nothing is routed. |
| `roe_ruleset_create` | Write | Creates a new, empty ruleset with its inputs. |
| `roe_ruleset_update` | Write | Renames a ruleset, changes its description, or adds and removes inputs. |
| `roe_ruleset_edit_rules` | Write | Changes rules in the draft: destination, condition, order, and the fallback rule. |
| `roe_ruleset_publish` | Write | Makes the draft live. This changes live routing. |
| `roe_ruleset_delete` | Write | Deletes a ruleset. Refused while a published, switched-on workflow uses it. |

> [accordion] How ruleset changes go live
>
> A ruleset has a draft and a live version. `roe_ruleset_create`, `roe_ruleset_update`, and
> `roe_ruleset_edit_rules` change the draft only. Leads keep routing by the live version until
> `roe_ruleset_publish`.
>
> Every write previews first and commits only on `confirm: true`. The preview shows each rule
> before and after, with queue and person names. The publish preview also shows changes to inputs
> and the fields they read.
>
> Before a publish, run `roe_ruleset_test` with `compare: true`. It runs each sample lead on the
> live version and on the draft and returns only the leads that would go somewhere else.
>
> `roe_ruleset_test` never assigns a lead, never writes to the evaluation log, and never moves a
> queue's order.

### Workflows

| Tool | Type | What it does |
| - | - | - |
| `list_workflows` | Read | Lists workflows with enabled state, published version, trigger types, and run counts. |
| `list_workflow_runs` | Read | Lists workflow runs, filtered by workflow, status, person email, company domain, trigger, or time. |
| `get_workflow_logs` | Read | Returns diagnostics, node visits, and a classified timeline for one workflow run. |
| `get_workflow_run_input` | Read | Returns the trigger payload that started one workflow run, to see what it was sent or repeat it. |
| `run_workflow` | Write | Starts a published workflow through its callable trigger. Effectful runs confirm in two steps. Turns on per workspace. |
| `workflow_describe` | Read | Lists the step and trigger types a workflow can be built from, and the fields each one takes. |
| `workflow_find_values` | Read | Lists what can go in one step input: earlier steps' outputs and the loader's options. |
| `workflow_read_graph` | Read | Reads a workflow's draft in slices: summary, nodes, edges or triggers, with a revision. |
| `workflow_validate_graph` | Read | Checks a draft the way a publish would, without changing anything. |
| `workflow_apply_changes` | Write | Edits a draft, or creates a new one, with one batch of changes. Previews first, then confirms. |
| `workflow_publish` | Write | Publishes a draft as the live version. Previews first, then confirms. |

For the build order, prompts that work, and rules to give your agent, see
[Build workflows with an AI agent](https://docs.os.default.com/mcp-workflow-building.md).

### Integrations & users

Read tools for checking whether your integrations and reps are set up correctly.

| Tool | Type | What it does |
| - | - | - |
| `integration_health` | Read | Reports whether each integration is `connected`, `broken`, or `not_connected`. |
| `mapping_coverage` | Read | Counts members with and without a CRM user mapping, per provider. |
| `mapping_validate` | Read | Checks a mapping value against the live provider to confirm the user ID actually exists. |
| `provider_people_list` | Read | Lists the users or owners that exist at a provider, with the value a mapping needs. |
| `calendar_connections_get` | Read | Reads a rep's connected calendars, primary calendar, and conferencing provider. |
| `users_list` | Read | Lists workspace members with role, status, and the user ID the per-user tools take. |
| `user_me` | Read | Reads your own member record: ID, email, name, and roles. |

### Identity graph

| Tool | Type | What it does |
| - | - | - |
| `identity_resolve` | Read | Resolves an email, domain, anonymous ID, or CRM record ID to its canonical identity and record. |
| `crm_duplicates_scan` | Read | Finds exact-key CRM duplicates: contacts sharing one email, or companies sharing one domain. |

### Data import & segments

Upload a CSV, map its columns to Default fields, and work with segments (saved table views).
This family turns on per workspace.

| Tool | Type | What it does |
| - | - | - |
| `upload_file` | Write | Starts a CSV upload. Returns a secure link a human opens to upload the file in their browser. |
| `get_upload_status` | Read | Polls an upload until the human completes it in their browser. |
| `upload_file_content` | Write | Uploads a small CSV inline in one call, for automated flows without a human. |
| `list_uploads` | Read | Lists CSV uploads with status and row and column counts. |
| `objects_map` | Write | Proposes a column-to-field mapping for a parsed CSV, with confidence and warnings. |
| `objects_update` | Write | Re-runs a proposed mapping with corrections before you confirm it. |
| `objects_confirm` | Write | Freezes a reviewed mapping proposal into an immutable confirmed version. |
| `segment_list` | Read | Lists the segments (saved table views) you can access, including CSV-backed ones. |
| `segment_get` | Read | Reads one segment's columns, saved filters, and data source. |
| `segment_query` | Read | Reads rows from a segment with pagination. The segment's saved filters always apply. |
| `segment_create` | Write | Creates a segment over data-model records or over an uploaded CSV. |
| `segment_update` | Write | Updates a segment's name, sharing, saved filters, or columns. |
| `segment_delete` | Write | Removes a segment definition. The underlying records or CSV are not touched. |

## Example prompts

Concrete, specific prompts help the client pick the right tools. Name the object, the fields,
and any filter, owner, or timeframe you care about.

**Explore what's there**

* "Which company fields can I filter and sort on?"
* "Show me the schema for person records, then list the 10 most recently created."

**Count and segment**

* "How many companies were created in the last 30 days?"
* "How many person records are missing a job title?"

**Query your CRM**

* "List Salesforce opportunities in the 'Negotiation' stage, sorted by amount, and page through them."
* "Count HubSpot companies in the technology industry with more than 500 employees."

**Filter by owner**

* "Find my Salesforce user ID, then list the accounts I own."
* "Look up Jordan Lee's HubSpot owner ID and count the deals they own."

**Run enrichment**

* "Enrich every person in my 'Inbound leads' view that's missing a company name."

**Inspect meetings and routing**

* “Rename my Q4 outbound booking link to Q4 EMEA outbound.”

* “Delete the revoked q4-promo link so I can reuse its identifier.”

* “Give me a booking link for the Sales intro event with me and Jordan Lee.”

* "List my upcoming meetings this week and who is hosting each one."

* "Why did this lead go to Jordan? Explain the assignment."

* "Who is up next in the 'Inbound AE' queue?"

* "What slots are open for the 'Intro call' event next week? Then create a booking link hosted by Jordan Lee that expires Friday."

**Check setup health**

* "Is our Salesforce integration healthy? Which reps are missing a CRM mapping?"
* "Does Jordan Lee have a connected calendar and working hours set?"

**Debug a workflow run**

* "Find the last failed run of the 'Inbound router' workflow and show me its logs."

**Build a workflow**

* "In the 'Inbound router' workflow, add a branch for companies with more than 1,000 employees
  that goes to the 'Enterprise AE' queue. Show me each change before you save it, then ask me
  before you publish."

**Do the math**

* "For companies in that view, what's the average deal size? Use the calculator for the exact number."

**Update a record (with confirmation)**

* "Set the lifecycle stage to 'Customer' on this HubSpot company. Show me the change to approve before writing."

## How access works

Default only exposes tools that your account is allowed to discover and call.

| Access layer | What it controls |
| - | - |
| Workspace enablement | Whether MCP access is available for your workspace. |
| Authentication | Whether the request comes from a Default user signed in through browser-based OAuth. |
| Role and grants | Admins and owners can use enabled tools. Members and editors can use booking tools, or routing writes with the required separate grants. |
| Event and queue access | Booking tools use existing app permissions. They grant no additional event or queue access. |
| Enabled features | Feature-gated tool families appear only while enabled for your workspace. |
| Connected integrations | Which external systems and records Default can use to answer (for example CRM queries). |

If a tool is not allowed for your account, it may not appear in your client's tool list.

For feature-gated tools, Default re-checks your workspace's current feature access on each tool
call. When a feature turns off, a client cannot use the cached tool.

### Give a member booking access

Every member and editor can use the booking tools. What they can book is bounded by their
existing access:

1. Give the person edit access to the required events through the existing event sharing controls.
2. Give the person access to any required queues, or add them as an active queue member.
3. Ask the person to refresh or reconnect their MCP client.

The booking tools can read permitted events, find people and visible queues, list booking links,
and create or revoke links. Each write still requires confirmation. Access checks also apply
when an agent supplies an event or queue ID directly.

Meeting reads show only meetings the person can see in the app. `meeting_cancel` previews
a meeting, then cancels it after confirmation if the person can manage it. Guests receive
the normal cancellation email. `meeting_mark_no_show` works the same way: the calendar event is
removed, the event's no-show messages are sent, and hosts picked from a queue get their spot back
when the queue's policy returns no-show meetings. `meeting_reschedule` moves a meeting to a new start
the way a host does from the calendar: minimum notice and working hours do not apply, and busy time is
shown as a warning only. `meeting_reassign` hands a meeting to another person or to a queue. A queue
picks its member when the change is confirmed, and the result names that person. The new host sends
a new calendar invite and the old invite is cancelled. `meeting_add_guests` adds up to 10 guests to a
meeting. A person who can manage the meeting can add guests even when the event does not let guests
invite others. `meeting_mark_held` marks a meeting as held once it has started. The calendar event
stays, no message is sent, and workflows that start when a meeting is held run. Held is final.
Booking tools do not enable personal availability settings, open-slot checks,
event editing, or workflow tools. Changing a person's role to viewer removes the booking tools
within 60 seconds.

## Troubleshooting

> [accordion] No tools appear after I connect
>
> Confirm that your workspace has MCP access enabled. Confirm that you are signed in with
> the correct account. Viewers see no booking tools, and routing writes need the required separate grants.
>
> Then refresh the tool list in your client. Available tools also depend on your connected
> integrations.

> [accordion] A newly enabled tool does not appear
>
> Refresh the MCP server in your client. Default sends the standard
> `notifications/tools/list_changed` notification when the client opens its update channel.
>
> Some clients keep an older tool cache or do not implement automatic list refresh. If a
> refresh does not send a new tool-list request, disconnect and reconnect the MCP server.

> [accordion] A listed tool says disabled or not found
>
> Your client likely cached a tool that is no longer enabled. Default enforces current feature
> access at execution time even when the client's display is stale.
>
> Refresh the MCP tool list to remove it.

> [accordion] My client says the server is unauthorized
>
> Sign in again to refresh your session.
>
> If the error continues, confirm that your client points at the Default MCP server URL. Also
> confirm that you completed the browser-based sign-in with the account that received access.

> [accordion] A tool appears but cannot answer my question
>
> Add more context to your prompt. Include the object, field, timeframe, owner, or filter you
> want the tool to use.
>
> If the data comes from an integration, confirm that integration is connected and synced.

> [accordion] A browser-based client gets an origin error
>
> Ask your Default admin or implementation team to confirm the client origin is allowed for
> your workspace's MCP configuration.

## For custom client builders

Connect your custom client to Default's remote MCP endpoint over Streamable HTTP. Then use
standard MCP requests to discover and call tools.

**1. Connect to the endpoint**

Connect to the Default MCP endpoint: `https://mcp.default.com/mcp`.

**2. Sign in with OAuth**

Complete the browser-based OAuth sign-in with your Default account. There is no access-token or
service-token flow.

**3. Initialize the session**

Send `initialize`. Then send `notifications/initialized`.

**4. Open the notification stream**

Open authenticated `GET /mcp` with `Accept: text/event-stream` to receive server
notifications.

**5. Discover the tools**

Call `tools/list` to discover available tools.

**6. Handle tool-list changes**

When `notifications/tools/list_changed` arrives, call `tools/list` again. Replace the cached
tool schemas.

**7. Call a tool**

Call `tools/call` with the selected tool name and arguments. Include the bearer credential
on every request.

Default's MCP endpoint does not require worker-affine session state.

> [!NOTE]
>
> Default's MCP server currently exposes tools. Resources and prompts may not be available from this
> endpoint unless your workspace has been explicitly configured for them.

```json Example tools/call request theme={null}
{
	"jsonrpc": "2.0",
	"id": 1,
	"method": "tools/call",
	"params": {
		"name": "data_count",
		"arguments": {
			"object": "company"
		}
	}
}
```

The response includes MCP content and, when available, structured output. Your client can
render the output as text, tables, or another UI.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.