MCP App Store
by KeenEthics
bitdrift icon

bitdrift MCP Server for ChatGPT

By bitdrift

Overview

Use bitdrift to investigate crashes and regressions in your mobile apps, inspect session logs, compare app versions, and create dashboards. Requires a bitdrift account with access to the relevant apps and telemetry. Sign in to bitdrift when connecting.

How to connect

  1. ChatGPT: Open the bitdrift listing in the ChatGPT app directory and choose Connect. ChatGPT listing bitdrift exposes 78 tools here.

Tools

bd_alerts_get

ChatGPT
Get alert configuration for a workflow chart. Requires alert read permission.

bd_alerts_history

ChatGPT
Get history for a workflow alert, newest first. Items omit details.alertConfig and details.contributingData unless includeDetails is set; page.history[].details.alertConfig is an object in discovery, use bd_schema for its full type. Continue with nextQuery. Requires alert read permission.

bd_alerts_list

ChatGPT
List workflow alerts using the public API filters and pagination. Requires alert read permission. Use bd_alerts_get for a selected configuration.

bd_alerts_test_notification

ChatGPT
Send a real test workflow or issue alert to the selected notification groups. Validates formatting and mentions. Requires alert write permission. Send only when explicitly requested.

bd_alerts_upsert

ChatGPT
Create or replace alert configuration for a workflow chart. Omitting alertConfig with an id deletes the alert. Requires alert write permission. Inspect bd_alerts_get before updating. Do not automatically retry an ambiguous creation result.

bd_apps_list

ChatGPT
List apps and versions observed in a time window (default: last 30 days), ranked by platform and app activity. Defaults to topApps=50 and versionsPerApp=20; appsLimited and versionsLimited report omissions. Build numbers are not provided by this API.

bd_capture_notifications_get

ChatGPT
Get notification settings for a workflow capture action. Use actionId from bd_workflows_get. Requires alert read permission.

bd_capture_notifications_upsert

ChatGPT
Replace notification settings for a workflow capture action. Requires alert write permission. Inspect existing settings with bd_capture_notifications_get before updating.

bd_charts_fetch_app_data

ChatGPT
Load the full workflow chart display payload for the MCP App.

bd_charts_list

ChatGPT
Discover accessible charts across workflows by workflow id, saved workflow view id, or partial chart name. SUMMARY is the default and returns ids, titles, kinds, and latest deployment status; FULL also returns the public chart metadata. Returns one native offset/limit page with totalCount. Pass nextQuery as the next call's arguments; budgetLimited marks a page shortened to fit. Use bd_charts_query for chart data.

bd_charts_query

ChatGPT
Query explicitly selected workflow metric, funnel, and Sankey charts. Discover chartRuleId, funnelRuleId, or sankeyRuleId with bd_workflows_get, or select charts on a custom dashboard by dashboardChartId from bd_dashboards_get to match the dashboard's display. Projection is required: SUMMARY returns time-series rollups only; CHART_DATA returns complete chart data, with line-chart buckets as timeSeries[].values arrays from timeSeries[].start (NaN marks a bucket with no value); FULL also includes supported metric detail. Metric charts use their configured display mode unless overridden. Defaults to the last 24 hours. Native limitStrategy.identifierMatch selects grouped series using returned IDs and full labels; histogramConfiguration selects percentiles. Unique-device counts are opt-in with SUMMARY or CHART_DATA; bucket rate details require CHART_DATA; FULL includes both. No pagination: narrow charts, time bounds, filters, or limits if the response exceeds the byte budget. Use sessions tools for capture charts.

bd_charts_render

ChatGPT
Open the workflow charts MCP App. Supply workflow_id and optional lookback, platform/app, top-K, sort, and percentile options; charts are discovered automatically. Defaults to a 24h lookback. Returns a small launch request; the App loads chart data for the user, and the model receives no chart values. Use bd_charts_query for any values you report or reason about.

bd_commands_get

ChatGPT
Get the latest state, progress, and result context for a device command. When succeeded.attachment is present, the result includes a resource link; small JPEG and PNG attachments are also returned as image content. Read the link with MCP resources/read to get the original bytes and MIME type, up to 8 MiB. Requires device debug permission.

bd_commands_history

ChatGPT
List retained direct device command submissions with caller identity, status, progress, and attachment metadata. Newest first; use pageToken to continue with unchanged filters. Does not include rejected submissions or SDK workflow executions. Requires device debug permission.

bd_commands_set_enabled

ChatGPT
Enable or disable a built-in or registered command. Enabling a privacy-sensitive built-in command requires agreeToPrivacySensitiveEnable=true. This command may collect personal or sensitive information. Before enabling it, make sure you have permission to collect this information and obtain user consent where required. Requires device command write permission.

bd_commands_unregister

ChatGPT
Remove a registered command. Returns a structured failure if workflows still reference it or it does not exist. Requires device command write permission.

bd_dashboards_delete

ChatGPT
Permanently delete the custom dashboard with the given id. Requires dashboard write permission. Inspect bd_dashboards_get before an ambiguous retry.

bd_dashboards_get

ChatGPT
Get a dashboard by ID. Custom dashboards return an outline by default: every tab, and per chart its id, chart identifier, title, layout, and backing workflow deployment status, plus heading text components. Read chart data as the dashboard shows it with bd_charts_query, selecting dashboard.dashboard.customDashboard.tabs[].charts[].id as dashboardChartId. List tab IDs in detailTabIds to return those tabs with full chart metadata and all text components; fetch every tab in detail before editing with bd_dashboards_upsert, which replaces the whole dashboard. When id is omitted, returns the legacy default-app health dashboard in full. Use bd_dashboards_list to discover IDs.

bd_dashboards_list

ChatGPT
List dashboards visible to the current user. SUMMARY is the default and returns compact records with id, title, owner, favorite, last-viewed, and edit status. FULL returns complete records; use bd_dashboards_get for one dashboard's chart and layout detail. Use pagination.limit and pagination.offset for bounded pages. total is the native matching count when the page contains results; an offset beyond the last page returns total 0.

bd_dashboards_upsert

ChatGPT
Create a custom dashboard with a generated ID when id is omitted. When id is set, update that dashboard if it exists or create one with the supplied ID if it does not; a typo can create a new dashboard. Supply all tabs and their layout components. Existing workflow charts can be referenced in tabs[].charts[].chartComponentLayout; inspect bd_schema for the nested authoring contract. Requires dashboard write permission. Fetch the dashboard before updating so existing components are preserved. Returns the dashboard ID and a per-tab summary, not the saved document; read it back with bd_dashboards_get. Do not automatically retry an ambiguous create.

bd_entities_get

ChatGPT
Retrieve an entity's sessions, activity, devices, issues, and recording status for the required timeRange. Select it by entityId, deviceId, or deviceCode. Device targets resolve their current entity when connected, otherwise their most recent entity in the time range; conflicting current associations require an explicit entityId. Current online devices are independent of historical observations. Reduce the time range or result limits if the response exceeds the output budget.

bd_entities_known_delete

ChatGPT
Delete a known entity by raw ID or hash. Requires entity write permission; deleting an absent entry returns not found. Check bd_entities_known_list before retrying an ambiguous result.

bd_entities_known_list

ChatGPT
List one offset-paginated page of known entities: app-user IDs, reported by the organization's own app through the bitdrift SDK, that its team bookmarked with a display name for debugging. Devices are omitted; use bd_entities_get for one entity's devices and sessions. Optionally filter by fuzzy name or recording status and sort by name, status, added, or viewed time. totalCount gives the match count; advance pagination.offset for another page and reduce pagination.limit if the response is too large.

bd_entities_known_upsert

ChatGPT
Create a known entity or replace its display name. Requires entity write permission. A supplied hash must identify an entity in the caller's recently viewed history. Inspect the resulting known entity before retrying an ambiguous result.

bd_entities_record_next_online_time_cancel

ChatGPT
Cancel an entity's pending next-online recording request. Requires entity write permission; repeated cancellation may return not found. Inspect the returned status before retrying an ambiguous result.

bd_entities_record_next_online_time_list

ChatGPT
List one offset-paginated page of recording requests for all visible entities or one entity. Filter by status and sort by entity ID, status, creation, or update time. totalCount gives the match count; advance pagination.offset for another page and reduce pagination.limit if the response is too large.

bd_entities_record_next_online_time_upsert

ChatGPT
Create or update an entity's pending next-online recording request. Replaces its notification group names when one is already pending. Requires entity write permission. Inspect the returned recording status before retrying an ambiguous result.

bd_issue_alerts_get

ChatGPT
Get crash, ANR, and exception alert configuration for an issues view. Requires alert read permission.

bd_issue_alerts_history

ChatGPT
Get firing history for a crash, ANR, or exception issue alert, oldest first. Items omit details.alertConfig unless includeDetails is set; page.history[].details.alertConfig is an object in discovery, use bd_schema for its full type. Continue with nextQuery. Requires alert read permission.

bd_issue_alerts_list

ChatGPT
List crash, ANR, and exception issue alerts using the public API filters and pagination. Requires alert read permission. Use bd_issue_alerts_get for a selected configuration.

bd_issue_alerts_upsert

ChatGPT
Create or replace crash, ANR, and exception alert configuration for an issues view. Omitting alertConfig deletes the view's alert. Requires alert write permission. Inspect bd_issue_alerts_get before updating. Do not automatically retry an ambiguous creation result.

bd_issue_groups_get_attributes

ChatGPT
Read all custom attributes for a crash, ANR, or exception issue group. Use bd_issue_groups_patch_attributes to upsert or remove keys. Requires issue read permission. No pagination or key filtering is supported; an oversized response returns a size error rather than truncating attributes.

bd_issue_groups_list

ChatGPT
Find and triage app crashes, ANRs, and exceptions, grouped into issue groups; start crash investigations here. Query one page of issue groups by time, app, version, platform, status, assignee, feature flags, and compound filters. Without viewId, defaults to the last 24 hours; returns 50 groups by default (maximum 100). Copy queryBounds into issueQuery.timeRange.fixedTimeRange and nextOffset into offset. viewId lists a saved issue-group view: fields set in issueQuery or issueGroupQuery replace its saved criteria, and its saved time range, or the last 7 days when it has none, applies unless timeRange is set, so queryBounds is absent. Pages may return fewer groups than limit to fit the response budget; budgetLimited marks this, and nextOffset continues from the first omitted group. Summaries omit event series and diagnostic details. Custom group attribute filters are not supported by the public API.

bd_issue_groups_list_assignees

ChatGPT
Discover users that can be assigned to crash, ANR, or exception issue groups, filtered by search or userIds. Returns one native page: offset defaults to 0, limit defaults to 100 (maximum 100). Advance offset by the number of returned assignees with unchanged filters; stop at total or an empty page. Requires issue assignment permission. Reduce limit or narrow filters if the response is too large.

bd_issue_groups_patch_attributes

ChatGPT
Apply ordered UPSERT and REMOVE operations to custom attributes of a crash, ANR, or exception issue group in one transaction. UPSERT requires a nonempty value; REMOVE requires an empty value. Unmentioned keys remain unchanged. Returns an empty object on success; use bd_issue_groups_get_attributes to inspect values. Requires issue assignment permission and records attribute change history. Do not automatically retry an ambiguous result.

bd_issue_groups_set_assignees

ChatGPT
Replace the complete assignee list for a crash, ANR, or exception issue group. Discover valid users with bd_issue_groups_list_assignees and pass their IDs in assignees. An omitted or empty assignees list clears assignment. Assigns users, not teams. Requires issue assignment permission.

bd_issue_groups_set_status

ChatGPT
Set the triage status of a crash, ANR, or exception issue group using its id from bd_issue_groups_list. Status values follow the public API; omitted status means NEW. FIXED records the latest observed app version and returns version or versionCode. No custom resolution version or snooze duration is supported. Requires issue assignment permission. Do not automatically retry an ambiguous result.

bd_issues_get

ChatGPT
Get the crash report for a single crash, ANR, or exception occurrence returned by bd_issues_list, including stack traces, optional thread details, and crash details. Returns issue identity, app information, session ID, processing status, and primary error stacks (or reporting-thread stacks when no error stacks exist). Optionally page through frame-free thread inventory with include.threadInventory=true, include.threadInventoryLimit, and include.threadInventoryOffset; follow threadInventory.nextOffset. Select reporting threads or stored IDs, and project frame components. Include fields, feature flags, app/device metrics, or binary images as needed. Optional section failures preserve core details. The 128 KiB result budget may shorten inventory pages with threadInventory.budgetLimited. Found issues include an MCP resource link in the tool result content for reading the complete unprojected report.

bd_issues_impact_summary

ChatGPT
Count users, sessions, and events affected by crashes, ANRs, and exceptions matching issueQuery and optional current issue group filters. Defaults to the last 7 days. Returns resolved fixed query bounds for the default window; provide an explicit fixed range to compare periods.

bd_issues_list

ChatGPT
Query one page of individual crash, ANR, or exception occurrences (issues) in an issue group, including session IDs. Defaults to the last 24 hours and 50 issues (maximum 100). Continue with nextPageToken and the same filters and sort; copy queryBounds into query.timeRange.fixedTimeRange and nextPageToken into tokenPagination.pageToken. Explicit relative ranges omit queryBounds; use a fixed range for stable pagination, since relative windows can shift. Does not hydrate details or automatically fetch subsequent pages.

bd_issues_trend

ChatGPT
Get crash, ANR, and exception occurrence counts over time for issues matching issueQuery and optional current issue group filters. Defaults to the last 7 days when issueQuery is omitted. The response includes the aggregation interval selected by the server.

bd_list_guides

ChatGPT
List bitdrift usage guides. Use this if you can't read MCP resources directly. Returns the complete catalog by stable ID, title, and summary in deterministic order.

bd_notification_groups_delete

ChatGPT
Delete a notification group by name. Requires notification group admin permission.

bd_notification_groups_list

ChatGPT
List notification groups and their configured targets. PagerDuty routing keys, DataDog API and application keys, and HTTPS webhook headers are omitted. Requires notification group read permission.

bd_notification_groups_test

ChatGPT
Send a real test notification to every target in the named group. Returns per-target delivery errors. Requires notification group admin permission. Send only when explicitly requested.

bd_read_guides

ChatGPT
Read bitdrift usage guides. Use this if you can't read MCP resources directly. Reads complete task guidance for IDs from bd_list_guides in request order. Unknown IDs fail the batch; request fewer guides when over budget.

bd_schema

ChatGPT
Inspect a mounted protobuf tool's input or output. Start with tool alone for the root, then use returned followups or canonical type names. Add expansionDepth for a bounded expansion of the root or selected message. Does not access account data.

bd_sessions_aggregate

ChatGPT
Aggregate numeric session log fields into time buckets using MEAN, MIN, MAX, or SUM. Supply explicit time bounds and a positive whole-millisecond aggregationWindow covering at most 2000 buckets; select up to 20 fields. Empty buckets are omitted. Prepares session data automatically when needed, waiting up to 30 seconds, but does not load records uploaded since the last read; use the first page of bd_sessions_logs for those. A pending result means the data becomes readable without action: wait pending.retryAfterMs milliseconds, then call pending.tool with pending.arguments; pending.reason says why. Errors mean waiting will not help: not_found explains why the session has no readable data, and failed_precondition reports a recent preparation failure; report these instead of retrying. If the response is too large, request fewer fields, a shorter time range, or a larger aggregationWindow.

bd_sessions_detail

ChatGPT
Read session metadata and optionally feature flag history or explicit as-of state. Flag keys can be discovered across the full session. As-of time must lie within the session. Feature flag failures preserve metadata in info and report featureFlagsError. Prepares session data automatically when needed, waiting up to 30 seconds, but does not load records uploaded since the last read; use the first page of bd_sessions_logs for those. A pending result means the data becomes readable without action: wait pending.retryAfterMs milliseconds, then call pending.tool with pending.arguments; pending.reason says why. Errors mean waiting will not help: not_found explains why the session has no readable data, and failed_precondition reports a recent preparation failure; report these instead of retrying.

bd_sessions_fetch_replay_frame

ChatGPT
Load the full session replay frame for the MCP App, preparing session data when needed.

bd_sessions_logs

ChatGPT
Query one page of session logs in timeline order; defaults to 50 records and the session time range. orQuery combines groups with OR; conditions within each group use AND. Empty filters add no constraints. Prepares session data automatically, waiting up to 30 seconds; the first page (no offset) also loads records uploaded since the last read. A pending result means the data becomes readable without action: wait pending.retryAfterMs milliseconds, then call pending.tool with pending.arguments; pending.reason says why. Errors mean waiting will not help: not_found explains why the session has no readable data, and failed_precondition reports a recent preparation failure; report these instead of retrying. Omit fields to exclude values; keys discovers this page's keys, selected returns named fields, and all returns all values. Continue with nextQuery as query and fields unchanged; absent nextQuery means done. Pages may return fewer rows than query.limit to fit the response budget; budgetLimited marks this, and nextQuery continues from the first omitted row. Select fewer fields to fit more rows per page; if a single row exceeds the budget, narrow fields.

bd_sessions_render_replay_frame

ChatGPT
Open a session replay frame in the MCP App. Supply session_id and optional raw timeline row_number; omit it for the latest valid frame. Returns a small launch request. The App loads the frame for the user, preparing session data when needed and loading new uploads for the latest frame; the model receives no frame contents, so use bd_sessions_logs for evidence you report. Oversized frames cannot render here; use the portal replay view.

bd_sessions_spans

ChatGPT
Search one page of session spans using native filters and page numbers. Default depth is zero and page size is 50 to bound results. Omit fields to exclude values, or request keys, selected, or all. Continue with nextQuery and the same fields selection. Prepares session data automatically, waiting up to 30 seconds; the first page also loads records uploaded since the last read. A pending result means the data becomes readable without action: wait pending.retryAfterMs milliseconds, then call pending.tool with pending.arguments; pending.reason says why. Errors mean waiting will not help: not_found explains why the session has no readable data, and failed_precondition reports a recent preparation failure; report these instead of retrying. Oversized pages are rejected, so retry with smaller perPage or depth.

bd_submit_feedback

ChatGPT
Report a problem with bitdrift tools to the bitdrift team: a missing capability, a failure, or a confusing description, schema, or result. Report once per task, after finishing or giving up, combining related problems. Do not include credentials, personal data, or account data values such as log contents or identifiers. Reports are stored with your account, user, and client name. Continue the task afterwards; do not retry a failed report.

bd_teams_delete

ChatGPT
Permanently delete a team by id. Requires teams write permission. Inspect bd_teams_get before an ambiguous retry.

bd_teams_get

ChatGPT
Get a team by id, including its name and description. Requires teams read permission. Use bd_teams_list to discover IDs.

bd_teams_list

ChatGPT
List teams visible to the current user. Filter by name or description and use pagination.limit and pagination.offset for bounded pages; default limit is 100. totalCount is the matching count when the page contains results. Requires teams read permission.

bd_teams_members_add

ChatGPT
Add organization users identified by userIds to a team. Existing memberships are preserved. Requires teams write permission.

bd_teams_members_remove

ChatGPT
Remove organization users identified by userIds from a team. Requires teams write permission.

bd_teams_upsert

ChatGPT
Create a team when id is omitted, or replace the name and description of the named team. An omitted description clears it on update. Requires teams write permission. Check bd_teams_list before retrying an ambiguous create.

bd_views_delete

ChatGPT
Permanently delete the saved view with the given id, including any issue alert configured on it. Requires view write permission. Inspect bd_views_get before an ambiguous retry.

bd_views_get

ChatGPT
Get one saved view with its filters, default sort, owner, and access. To list what a view matches, pass its id as viewId to bd_issue_groups_list or bd_workflows_list. Requires view read permission.

bd_views_list

ChatGPT
Discover saved issue-group and workflow view IDs and names. Returns one page of summaries, default 50; filters, default sort, and access are omitted, so use bd_views_get for them. Set serviceType to ISSUE_GROUP or WORKFLOWS and name to narrow. Issue alerts target an issue-group view id. Pass nextQuery as the next call's arguments; absent means done. Pages may return fewer views than the limit to fit the response budget; budgetLimited marks this, and nextQuery continues from the first omitted view. Requires view read permission.

bd_views_upsert

ChatGPT
Create a saved issue-group or workflow view with insert, or change one with update. Issue-group views use viewFilters.issueGroupListNew. Inspect bd_schema for the filter contract. An update changes only supplied fields; a supplied viewFilters replaces all saved filters, so fetch bd_views_get first and resend every filter to keep. defaultSort entries must match the view type. Requires view write permission. Do not automatically retry an ambiguous insert.

bd_whoami

ChatGPT
Return the authenticated user email and tenant name for this connection.

bd_workflows_chart_metadata_upsert

ChatGPT
Replace the metadata for one workflow rule's chart, including title and display configuration. Discover workflowId and ruleId with bd_workflows_get; inspect bd_schema for the metadata contract. Requires workflow write permission.

bd_workflows_create

ChatGPT
Create a validated, undeployed workflow from an inline workflow definition, optional metadata, and per-rule chart metadata. Inspect bd_schema with tool bd_workflows_create, then follow its nested type references to discover the full authoring contract. Requires workflow write permission. Returns the new id; use bd_workflows_deploy with that id as workflowId to activate it. Creation never deploys. Do not automatically retry an ambiguous failure: creation is not idempotent and could create duplicates. Inspect bd_workflows_list before retrying.

bd_workflows_delete

ChatGPT
Permanently delete a workflow by workflowId. This is irreversible and requires workflow write permission plus ownership or workflow admin permission. Nonempty violations mean deletion failed; resolve them before retrying. An empty violations list indicates success.

bd_workflows_deploy

ChatGPT
Deploy an existing workflow by workflowId, making it active and recording a new immutable deployment version. Requires workflow deploy permission and deployment access to the workflow. Nonempty violations mean deployment failed; inspect them before retrying. An empty violations list indicates success. Workflow and chart metadata are omitted; use bd_workflows_get when you need the resulting definition or state. Do not automatically retry an ambiguous result; inspect bd_workflows_get first.

bd_workflows_deployment_history

ChatGPT
List historical workflow deployments visible to the current user. Defaults to summaries with workflow id/name, version, and deploy/stop times; FULL includes public deployed workflow definitions, excluding internal client configurations. Filter by workflowId and use offset and limit for native pagination. Pass nextQuery as the next call's arguments; budgetLimited marks a page shortened to fit. For large definitions, select one workflow and a small limit.

bd_workflows_get

ChatGPT
Retrieve a workflow definition, deployment state, chart identifiers, and capture actions. Use detail.workflow.actions[].flushRule.id as sessions query.actionId; use charts.charts[].id.workflow.chartRuleId as bd_charts_query charts[].chartRuleId. chartsStatus reports INCLUDED, FAILED, or TOO_LARGE; chartsError supplies details when space permits. Definitions are not paginated; oversized definitions return a size error.

bd_workflows_list

ChatGPT
Discover workflow IDs, names, deployment state, and metadata. Returns one page, default 50; flows and actions are omitted. Public query takes precedence over filters and platformTargets. viewId lists a saved workflow view: fields set in query or platformTargets replace its saved criteria. To discover Instant Insights workflows, pass {"query":{"tags":{"anyOf":[{"conditions":[{"operator":"INCLUDES","match":"ANY_OF","tags":["instant-insight"]}]}]}}}. For query.ootbEventTypes, call bd_schema with tool bd_workflows_list, direction INPUT, and typeQuery OotbConditionType; inspect the matching enum types for event names and documentation. Each filter entry selects one of generic, android, or apple. Pass nextQuery as the next call's arguments; absent means done. Use bd_workflows_get for definitions and chart/capture identifiers. Pages may return fewer workflows than limit to fit the response budget; budgetLimited marks this, and nextQuery continues from the first omitted workflow.

bd_workflows_rename

ChatGPT
Rename one workflow without changing its definition, metadata, tags, or deployment. Requires workflow write permission and edit access. Returns the updated name and id; use bd_workflows_get to inspect the full workflow.

bd_workflows_sessions

ChatGPT
Return one page of sessions captured by this workflow, default 50 over the last 24 hours. Omit query.actionId only when the workflow has one capture action; otherwise choose from the reported IDs. orFilter groups use OR, with AND inside each group. Omit fields to exclude values; keys discovers this page's keys, selected returns named values, all returns all. Pass nextQuery as query with the same workflowId and fields; absent means done. Explicit relative windows retain API semantics; use fixed bounds for stable paging. Oversized results: reduce query.perPage or narrow fields and filters.

bd_workflows_stop

ChatGPT
Stop a deployed workflow by workflowId, making it idle. Requires workflow deploy permission and deployment access to the workflow. Returns an empty object on success; workflow and metadata are omitted. Use bd_workflows_get to inspect the resulting state before retrying an ambiguous result.

bd_workflows_tags_list

ChatGPT
List the available workflow tags. Use these values to filter bd_workflows_list or assign tags with bd_workflows_tags_set. This endpoint does not paginate; a large tag set can exceed the response budget.

bd_workflows_tags_set

ChatGPT
Replace the complete set of tags assigned to one workflow. Pass an empty tags array to remove all tags. Requires workflow write permission and edit access. Returns the resulting tags.

bd_workflows_update

ChatGPT
Replace an existing workflow configuration by workflowId with the entire updated definition and optional metadata. Omitting metadata clears the existing metadata, including the description, rule titles, and deployment expiration settings. Fetch bd_workflows_get and resend any metadata you want to preserve. Requires workflow write permission and edit access. Inspect bd_schema with tool bd_workflows_update for the authoring contract. Nonempty violations mean the update failed; resolve them before retrying. Workflow and metadata are omitted from the response; use bd_workflows_get to inspect the resulting definition. Updating does not deploy the workflow. A deployed workflow accepts metadata-only updates; to change its configuration, stop it with bd_workflows_stop, update it, and deploy it again with bd_workflows_deploy. Inspect bd_workflows_get before retrying an ambiguous result.

Capabilities

WritesInteractive

Example Prompts

Click any prompt to copy it.

Similar Apps

Listed on MCP App Store

Listed on MCP App Store

Add this badge to your site or README. It links to this page.

HTML

<a href="https://mcp-app-store.keenethics-labs.com/app/bitdrift"><img src="https://mcp-app-store.keenethics-labs.com/badges/listed-on-mcp-app-store.svg" alt="Listed on MCP App Store" width="220" height="44"></a>

Markdown

[![Listed on MCP App Store](https://mcp-app-store.keenethics-labs.com/badges/listed-on-mcp-app-store.svg)](https://mcp-app-store.keenethics-labs.com/app/bitdrift)