Lightmeter

By Lightmeter Ltd

Overview

Manage Lightmeter outbound campaign workspaces from ChatGPT, including campaigns, leads, email sequences, campaign creative drafts and approvals, Unibox replies, performance analytics, and deliverability insights.

Tools

campaign_activation_readiness_get

ChatGPT

Evaluate the same campaign activation blockers and readiness state shown in the browser for one visible campaign.

campaign_analytics_leads_list

ChatGPT

List per-lead campaign analytics and refs for drilling into individual lead activity.

campaign_analytics_lifetime_get

ChatGPT

Get lifetime totals, rates, rate context, and drilldowns for one campaign.

campaign_analytics_sequence_get

ChatGPT

Get date-grained sequence-email performance for finding which steps or variants drive results.

campaign_analytics_timeseries_get

ChatGPT

Get date-grained campaign analytics for trend analysis across a requested period.

campaign_audience_import_apply

ChatGPT

Add audience leads to a customer-owned campaign under its dedupe policy; an optional dedupe scope updates the campaign policy first.

campaign_audience_import_preview

ChatGPT

Preview the dedupe, blocklist, and send eligibility impact of importing a customer-owned audience into a campaign, including an optional proposed dedupe policy.

campaign_creative_approve

ChatGPT

Explicitly approve the current Creative draft and materialize it into campaign steps.

campaign_creative_approve

ChatGPT

Approve the current Default Creative as one immutable snapshot and mark it Ready for future campaign sends. When campaign_creative_get reports its server-owned adaptations_activated capability as true, the snapshot also includes every active sparse adaptation. Inactive adaptation history is excluded from the snapshot but remains visible through get.

campaign_creative_checkpoint_create

ChatGPT

Save a manual checkpoint for every current Creative document. Does not approve or send.

campaign_creative_checkpoint_restore

ChatGPT

Restore one Creative document checkpoint when the expected workspace revision still matches. Does not approve or send.

campaign_creative_comment_add

ChatGPT

Add an agent-authored comment thread to a Creative document in the selected campaign.

campaign_creative_comment_delete

ChatGPT

Soft-delete a Creative comment thread in the selected campaign. The thread is removed from the active workspace history and a deletion event is recorded.

campaign_creative_comment_reopen

ChatGPT

Reopen a resolved Creative comment thread.

campaign_creative_comment_reply

ChatGPT

Add an agent-authored reply to an existing Creative comment thread.

campaign_creative_comment_resolve

ChatGPT

Mark a Creative comment thread resolved.

campaign_creative_generate

ChatGPT

Generate or regenerate a transient Creative draft using the same quotas and model contract as the UI. This consumes included AI generation quota. It does not save, approve, or send Creative. Supply the revision from campaign_creative_generation_capability_get and a fresh operation UUID. A completed operation cannot replay its transient preview.

campaign_creative_generated_draft_apply

ChatGPT

Replace the editable Creative working copy with a successful transient generated draft when its base revision still matches. Creates a recovery snapshot. Does not approve or send.

campaign_creative_generation_capability_get

ChatGPT

Read the current Creative revision, model, input limits, and workspace/campaign generation quota before requesting a draft.

campaign_creative_get

ChatGPT

Get the current editable Creative draft documents for a campaign.

campaign_creative_get

ChatGPT

Get the current editable Default Creative plus every target-labelled sparse Google, Microsoft, and SEG adaptation document, including inactive history. The response's server-reported adaptations_activated value is a capability signal, not an input or filter. Non-Default actions are available only when it is true; false means Default-only writes.

campaign_creative_snapshot_restore

ChatGPT

Restore a retained Creative recovery snapshot only when the expected current revision still matches. Creates a pre-restore snapshot and does not approve or send.

campaign_creative_snapshots_list

ChatGPT

List up to three retained recovery snapshots created before generated Creative replacement or a prior restore. Snapshot payloads are not exposed.

campaign_creative_update

ChatGPT

Write subject/body text into the editable Creative draft. This creates draft history but does not approve.

campaign_creative_update

ChatGPT

Write Default subject/body text into the editable Creative working copy. Call campaign_creative_get first: its server-reported adaptations_activated capability must be true before submitting a non-Default adaptation_target or adaptation_active. Omitting adaptation_target makes only that submitted field Default; it does not hide existing adaptations from get. Sparse adaptations inherit missing fields from Default. adaptation_active=false performs a recoverable reset, true explicitly reactivates, and omission preserves an existing adaptation's lifecycle state. This creates history but does not approve.

campaign_creative_workspace_get

ChatGPT

Read editable Creative documents, checkpoint history, comment threads, and the current revision. Live browser collaboration sessions are not returned.

campaign_delete

ChatGPT

Soft-delete an ended or draft campaign. Requires confirm=true. Active, paused, or ending campaigns must be ended first.

campaign_duplicate

ChatGPT

Copy reusable campaign content and settings into an unsent draft. Defaults to no mailboxes or routing. To share eligible source infrastructure, use campaign_infrastructure_sharing_preview, explain its reputation/capacity warnings, and provide share_infrastructure=true, the current fingerprint and risk acknowledgement. Sharing permits concurrency; no automatic sequential handoff or launch. Queues normal configured campaign-created webhooks after commit.

campaign_email_preview

ChatGPT

Render submitted templates using current campaign click tracking. Enabled tracking returns already-working anonymous redirects, reused for identical domain/destination pairs. Does not save Creative, send email or collect engagement. An optional owned sending_domain selects the redirect host; otherwise prefer an assigned domain or an existing ready workspace domain.

campaign_end

ChatGPT

Deliberately stop an active or paused campaign. The command returns while ending is in progress; poll campaign_lifecycle_get until status is ended. Reuse the same operation_id when retrying.

campaign_fast_forward

ChatGPT

Send at most the next queued message to each eligible lead, then complete the campaign. Available only while sending is enabled and campaign completion is at least 95%. This cannot be undone.

campaign_follow_up_create

ChatGPT

Create an unsent follow-up draft from selected eligible completed contacts, retaining source membership. Defaults to no infrastructure; sharing requires the current review fingerprint and risk acknowledgement. Does not launch or resume sending. Queues normal contact-analysis jobs after commit.

campaign_follow_up_preview

ChatGPT

Read selected completed contacts without replies eligible for a follow-up. No campaign or membership changes.

campaign_infrastructure_sharing_preview

ChatGPT

Read eligible workspace mailboxes, their owning campaigns, and current review fingerprint. Returns reputation/capacity warnings and concurrent/manual-sequential behavior. No assignments or campaigns change.

campaign_lead_activity_list

ChatGPT

List chronological campaign activity for one lead after selecting a lead from campaign analytics.

campaign_lead_personalization_update

ChatGPT

Update a campaign lead's email, name, company, or custom attributes using the browser's identity and validation rules.

campaign_lead_quality_diagnostics_get

ChatGPT

Get existing lead-quality diagnostics for a campaign, including email verification counts and explicit unavailable signals.

campaign_lead_removal_reasons_list

ChatGPT

List active reason codes, labels, and text requirements for removing leads from this campaign.

campaign_lead_remove

ChatGPT

Remove one campaign lead with a recorded reason and confirm_removal=true.

campaign_leads_add

ChatGPT

Add leads to an existing Lightmeter campaign.

campaign_leads_list

ChatGPT

Search, filter, and page the campaign's current leads with the same fields and status counts as the browser. Page size is at most 100; reduce it if the 128 KiB result limit is reached.

campaign_leads_move

ChatGPT

Move the confirmed selected leads into an existing campaign or a new cloned/blank unsent draft. Removes source membership. A new destination defaults to no infrastructure; explicit sharing requires the current review fingerprint and risk acknowledgement. Existing destinations cannot acquire infrastructure through a lead move. Creating a destination queues normal contact-analysis jobs after commit.

campaign_leads_move_preview

ChatGPT

Read the number of selected leads that can be moved. No membership or campaign changes.

campaign_leads_remove

ChatGPT

Remove the selected campaign leads using a removal reason and confirm_removal=true. Preview the same selection first. This action affects campaign lead membership and history and cannot be undone by this tool. Large filtered results return removed_count with removed_lead_ids_omitted instead of the full ID list.

campaign_leads_remove_preview

ChatGPT

Count selected, removable, and nonremovable leads before a bulk removal. No data is changed.

campaign_leads_top_up_csv

ChatGPT

Preview or apply a CSV top-up. Preview first; if shared-lead changes require review, use its review_token and an explicit preserve or overwrite decision. CSV arguments are limited to 1 MiB and results to 128 KiB. Page large preview rows with review_offset and review_limit; large apply results retain summary counts and omit per-row details available through the public API.

campaign_lifecycle_get

ChatGPT

Get the canonical campaign status, pause reason, identity-aware available actions, sendability/readiness blockers, and any pending lifecycle operation.

campaign_link_tracking_update

ChatGPT

Turn click tracking on or off for one campaign without changing the workspace default. Existing redirects and recorded clicks remain available; prepared message bytes are preserved.

campaign_mailbox_allocation_policy_get

ChatGPT

Read the same customer-visible overlap-allowed flag shown in the browser; detailed staff routing remains private.

campaign_mailboxes_assign

ChatGPT

Replace the complete set of active campaign mailbox assignments with the supplied workspace mailboxes. Include every current assignment you want to keep; omitted mailboxes are removed, and an empty list clears assignments subject to existing continuity safeguards. To deliberately share occupied domains, preview first and provide allow_shared_infrastructure=true, the current fingerprint and risk acknowledgement. Existing sending limits and lead continuity remain enforced. Campaign-wide assignment is unavailable while mailbox overlap is enabled; use the sending-pool add operation.

campaign_pause

ChatGPT

Pause an active campaign. Reuse the same operation_id when retrying the same command.

campaign_paused_lead_recovery_apply

ChatGPT

Apply an unchanged reviewed paused-lead recovery plan atomically. The campaign's Sending (lifecycle status active) or Paused status is preserved; resuming a Paused campaign is a separate command.

campaign_paused_lead_recovery_defer

ChatGPT

Explicitly defer the current mailbox-change paused-lead recovery obligation with a reason. The exact obligation ID and revision must come from the lifecycle or recovery preview. This preserves the campaign's Sending (lifecycle status active) or Paused status; resuming a Paused campaign is a separate command.

campaign_paused_lead_recovery_preview

ChatGPT

Preview the all-or-nothing paused-lead recovery plan for a Sending (lifecycle status active) or Paused campaign. Sending recovery is limited to the current mailbox-change obligation. For leads that already received campaign emails, supply an explicit previous-mailbox to replacement-mailbox mapping. Inspect continuity_requirements for affected-lead and remaining-campaign-email counts, required recipient mailbox-host sending pools, eligible replacement mailboxes, and exact ineligibility reasons.

campaign_reopen

ChatGPT

Move an ended or completed campaign into paused review state without changing lead states, mailbox assignments, history, or results. This never resumes sending.

campaign_replies_list

ChatGPT

List qualified human reply excerpts and classifications for a campaign/date range when diagnosing why performance changed.

campaign_restore

ChatGPT

Restore a soft-deleted campaign using the same ownership and domain checks as the browser.

campaign_resume

ChatGPT

Resume a paused campaign after all readiness checks pass. If newer working Creative is not marked Ready, the command requires explicit acknowledgement that those newer changes will not be sent. This action is available to authenticated workspace users.

campaign_schedule_update

ChatGPT

Set the campaign's IANA timezone and weekly sending windows using the same customer-scoped schedule operation as the browser.

campaign_send_timeline_get

ChatGPT

Get the same campaign send timeline projection shown in the campaign editor.

campaign_sending_pool_mailboxes_add

ChatGPT

Atomically assign workspace mailboxes and add them to the named boxhost target's sending pool using its current revision. Preview sharing first to use occupied sending domains, then explicitly acknowledge reputation and shared-capacity risks. Preserves source campaign assignments and started-lead continuity; does not change sending limits or launch sending.

campaign_sending_pool_mailboxes_list

ChatGPT

Read mailbox candidates, shared-add availability and current routing revision for one boxhost target. Does not change assignment or pool membership.

campaign_sending_pools_get

ChatGPT

Get the sending pools and assigned mailboxes for a campaign in the authorized workspace, including the current revision for later pool updates. Use this to inspect how the campaign is configured to send; does not change assignments, pool membership, or routing.

campaign_sequence_reorderability_get

ChatGPT

Inspect whether saved campaign steps may be reordered and their current order.

campaign_sequences_list

ChatGPT

List saved campaign steps, variants, and current subject lines visible to the workspace.

campaign_subject_pause_set

ChatGPT

Enable or disable one current subject option without editing or reapproving the shared email body. At least one subject per body must remain enabled.

campaign_test_email_send

ChatGPT

Send at most 100 test messages per request from an active physical mailbox assigned to this campaign and workspace to 1-50 named test addresses. Requires confirm_send=true; variant_label requires step_number. This makes real outbound sends.

campaign_visit_history_get

ChatGPT

Read accepted tracked-send counts, distinct original leads, rates and step/variant/subject attribution in an explicit retained scope, including deleted campaigns. Never recreates a lead or authorizes a new send.

campaigns_get

ChatGPT

Get one campaign by ID from a workspace when campaign metadata or navigation links are needed.

campaigns_list

ChatGPT

List campaigns in a workspace and return refs for follow-up campaign detail or analytics tools.

campaigns_submit

ChatGPT

Create and submit a campaign in a Lightmeter workspace.

capacity_allocation_get

ChatGPT

Get how the workspace's daily sending capacity is allocated, as shown in Infra › Allocation: subscribed, usable, in campaigns, available, reserve (labelled backups), recovering and warm-up per day; each non-ended campaign's capacity, recent sends, leads and projected completion; what needs attention; the last applied capacity plan; and can_apply_plans. Set include_days=true for the daily history and projection series from history_from (default 14 days ago, at most 180 days back). Lightmeter staff also receive the labelling dry run.

capacity_labelling_apply

ChatGPT

Lightmeter staff only. Apply suggestions from capacity_labelling_review_get by id: each selected mailbox or whole domain gets its suggested label through the same lifecycle status change staff make on the Mailboxes page, recorded in Activity with the reason. The review is recomputed first; selected suggestions that are no longer suggested are skipped with reason no_longer_suggested, and a suggestion whose mailbox status changed while applying is skipped with mailbox_status_refused. A domain change applies to every mailbox of the domain or none. Mailboxes in campaigns are never changed. Returns applied_ids, mailboxes_relabelled, skipped_count and skipped (at most 500 listed).

capacity_labelling_review_get

ChatGPT

Lightmeter staff only. List what automatic labeling of mailboxes not in a campaign would change now, as shown in Infra › Allocation › Review and apply: Available mailboxes to Recovering (unhealthy for every target for 2 checks running), Recovering mailboxes back to Available (healthy for Google or Microsoft for 2 checks running), and whole Available domains to Backup or surplus Backup domains back to Available to keep backups at the Subscribed target. Each suggestion has an id, the mailbox or domain, from_status and to_status (lifecycle codes: not_in_use is Available), per_day capacity and the reason. Returns at most limit suggestions (default and maximum 200) with total_suggestions; counts always cover every suggestion. Nothing changes; apply with capacity_labelling_apply.

capacity_plan_apply

ChatGPT

Lightmeter staff only. Apply a capacity plan reviewed with capacity_plan_preview: assign the planned whole domains to each campaign and its sending pools, and remove returned domains from completed, ended or paused campaigns (started leads stay on their mailboxes or are reassigned, as chosen). Pass the same campaigns and returns plus the preview's fingerprint; if capacity changed since the preview, nothing changes and the new preview is returned. Recorded in Activity.

capacity_plan_preview

ChatGPT

Preview a capacity plan without changing anything: which whole Available domains (or reserve domains when use_reserve=true) would be added to each campaign to reach a target, and which domains completed, ended or paused campaigns would return. Returns per-campaign before/after sends per day (before counts a shared domain once, under the campaign that is sending, as Allocation does), the domains, projected completion before and after, any shortfall and its reason, and the plan fingerprint. Only Lightmeter staff can apply plans.

instant_inbox_test_get

ChatGPT

Lightmeter staff only. Get one workspace test, including per-seed results and the pending_set_fingerprint for mark-missing.

instant_inbox_test_mailboxes_list

ChatGPT

Lightmeter staff only. List workspace mailboxes with eligibility and route readiness. Supports search, include_hidden, cursor, sort_by and sort_direction.

instant_inbox_test_mark_missing

ChatGPT

Lightmeter staff only. Requires write scope. Finalize the exact pending set using pending_set_fingerprint from test detail as If-Match and a stable idempotency_key. Grid decides eligibility and distinguishes Missing from observation errors. Reuse the key on retries.

instant_inbox_test_preflight

ChatGPT

Lightmeter staff only. Review the exact message and seed panel without sending mail. Requires write scope. Preflight is not capped. Supply a stable operation_id; retain it and the returned review_seal for start. For an already-started operation_id, preflight returns idempotent_replay=true and existing_test_id with no review_seal: use instant_inbox_test_get on that ID, do not call start.

instant_inbox_test_start

ChatGPT

Lightmeter staff only. Requires write scope. Sends real mail to the seed panel using the exact message reviewed by preflight. Grid enforces a 60-tests-per-UTC-day cap across all workspaces; preflight is not capped. Once 60 tests have started in the current UTC day, new starts return HTTP 429 daily_test_limit_reached. Run one test per exact message. Starting is non-idempotent without a key: operation_id is required and forwarded as Idempotency-Key. Reuse the same operation_id and review_seal on retries; do not create a new operation to retry a send. On HTTP 409 review_required, re-run preflight with the SAME operation_id, then start with the new seal.

instant_inbox_tests_list

ChatGPT

Lightmeter staff only. List workspace tests with cursor pagination.

onboarding_industry_keywords_suggest

ChatGPT

Suggest industry keywords from a public customer website using a Setup invitation linked privately to your signed-in user. Supply access_id and payload to generate suggestions, or call with empty arguments to list linked invitations and open the private linking page. Fetches the website and uses generation allowance, but saves no keywords and starts no onboarding work. Representative acceptance-only links cannot perform Setup. Manual entry remains available when status is unavailable.

onboarding_native_agreement_accept

ChatGPT

Complete a binding standard agreement on the customer's behalf only after the human explicitly accepts the exact server acceptance_statement. Supply agreement_id, a stable UUID request_id, verification_token, accepting name, confirmed customer_address, notice_email and authorized=true in payload. Acceptance alone completes it; payment is separate. Never automatically reissue after an unknown result: use acceptance_status.

onboarding_native_agreement_acceptance_status

ChatGPT

Resolve an unknown acceptance result without accepting. Supply original agreement_id and request_id. A definite_non_acceptance answer fences that UUID; a human's intentional retry must use a new UUID. The same recovery call is safe to repeat.

onboarding_native_agreement_copies_retry

ChatGPT

Explicitly retry failed or unknown copies of the immutable accepted PDF to the verified acceptor and invoices@lightmeter.io. An unknown SMTP result can duplicate a copy; contract formation never repeats.

onboarding_native_agreement_delegate

ChatGPT

Send a separate acceptance-only link to another person. Supply agreement_id, name and their email in payload. This link cannot pay or set up the workspace.

onboarding_native_agreement_email_verify

ChatGPT

Check whether the accepting email was verified privately in the browser. Never ask for an email code or verification token in chat. Open the returned verification page if verification is needed.

onboarding_native_agreement_get

ChatGPT

Read the current immutable native offer and exact acceptance statement using the private invitation capability. A Beam workspace login alone does not grant this capability.

onboarding_native_agreement_payment_refresh

ChatGPT

Read authoritative Stripe confirmation for the reserved Checkout. Browser returns cannot confirm payment. Representative or existing-customer agreement-only links cannot reconcile payment.

onboarding_native_agreement_pdf_chunk_get

ChatGPT

Read exact retained issued or accepted PDF bytes by chunk, using the private invitation capability. Verify the returned SHA256 after assembling all chunks.

onboarding_native_agreement_verification_code_send

ChatGPT

Send a verification code to the accepting person's own email address. Supply agreement_id and email in payload.

sending_pool_swap_apply

ChatGPT

Lightmeter staff only. After staff confirmation, apply the exact reviewed sending_pool_swap_preview. Removes only the specified provider memberships and immediately moves their started leads using swap. Other pools stay unchanged.

sending_pool_swap_preview

ChatGPT

Lightmeter staff only. Review a swap flag from capacity_allocation_get. Nothing changes. Uses the existing replacement planner, Available before reserve.

unibox_compose_leads_list

ChatGPT

Find existing workspace leads by name or email before composing a message. Only a stored lead can receive a new Unibox message.

unibox_compose_sent_steps_list

ChatGPT

List the campaign steps actually sent to a stored lead, including their message and thread IDs for reply or new-message context.

unibox_lead_compose_context_prepare

ChatGPT

Create or find the existing lead's canonical Unibox conversation for a sent campaign step without sending email. The returned thread UUID can be used to stage attachments; expected_latest_message_id guards a subsequent send.

unibox_lead_message_send

ChatGPT

Send a plaintext reply to a campaign step already sent to an existing lead, even if the lead has never replied. Set reply_parent_message_id to the selected step's message_id. For a separate unthreaded conversation, set new_message, context_message_id, and subject. The selected step determines the sender. This sends live email.

unibox_observed_message_content_get

ChatGPT

Read the current provider body for one accepted CRM message. First get thread detail and use its observation_id and observation_revision. Canonical messages require observed_event_id; historical Grid threads require source_kind grid and no event ID. A changed source requires reloading the thread. Sanitized HTML and plain text are limited to 1 MiB combined; this read does not change reply admission.

unibox_reply_attachment_delete

ChatGPT

Delete one unsent attachment draft owned by this MCP user and thread. Sent attachments cannot be deleted here.

unibox_reply_attachment_read

ChatGPT

Read a sent workspace attachment as bounded 32 KiB base64 chunks. Pass content_sha256 on continuation to reject changed content; drafts cannot be read.

unibox_reply_attachment_recent_list

ChatGPT

List up to five confirmed reusable reply attachments in an accessible workspace thread.

unibox_reply_attachment_reuse

ChatGPT

Copy one confirmed recent attachment into a 24-hour draft owned by this MCP user and thread. This does not send mail.

unibox_reply_attachment_stage

ChatGPT

Stage a file of up to 1 MiB for a reply in an accessible Unibox thread. The draft expires after 24 hours; staging does not send mail.

unibox_thread_classification_impact_get

ChatGPT

Preview the credit impact of a proposed reply category without changing the thread classification or billing state.

unibox_thread_forward_send

ChatGPT

Send a live forward to explicitly named recipients using a workspace sender mailbox. Uses a committed idempotency claim before delivery.

unibox_thread_get

ChatGPT

Get one canonical UniBox thread with chronological messages and classification evidence.

unibox_thread_lead_domain_blocklist

ChatGPT

Add the lead email domain represented by a workspace Unibox thread to the workspace blocklist.

unibox_thread_lead_email_blocklist

ChatGPT

Add the lead email represented by a workspace Unibox thread to the workspace blocklist.

unibox_thread_lead_unsubscribe

ChatGPT

Unsubscribe the campaign lead represented by a workspace Unibox thread. This also records activity and blocklists the lead email.

unibox_thread_reply_personalize

ChatGPT

Preview Unibox reply template fields in text and HTML using the thread's lead and sender context. At most 64 template expressions are accepted; oversized substituted values or estimated rendered content are rejected. This does not send mail.

unibox_thread_reply_send

ChatGPT

Send a live plaintext reply to a canonical UniBox thread after checking the latest message id.

unibox_threads_list

ChatGPT

List canonical UniBox threads in a workspace, including unanswered interested replies and current classification evidence.

workspace_activity_events_list

ChatGPT

Read the workspace activity log, with a maximum of 100 events per page and 20 values per filter list.

workspace_agreement_pdf_chunk_get

ChatGPT

Read one bounded chunk of a customer-owned issued or executed agreement PDF. Repeat reads are safe; every call requires current workspace access. Assemble chunks in index order and verify the returned SHA-256 digest and total byte length.

workspace_agreements_list

ChatGPT

Read customer-visible agreement metadata for a workspace without changing legal state.

workspace_analytics_overview_get

ChatGPT

Get workspace-level analytics overview for a requested date range.

workspace_api_key_create

ChatGPT

Create a workspace API key and return its token once. Requires Admin or Owner; treat the result as private.

workspace_api_key_delete

ChatGPT

Permanently delete a workspace API key. Requires Admin or Owner.

workspace_api_key_update

ChatGPT

Replace the name or expiration settings of a workspace API key. Requires Admin or Owner.

workspace_api_keys_list

ChatGPT

Read workspace API-key metadata when the signed-in member can manage API keys. Token values are never returned.

workspace_audience_create

ChatGPT

Create a named static audience from 1 to 2000 existing active workspace lead UUIDs.

workspace_audience_csv_apply

ChatGPT

Atomically create an audience and import its CSV leads using the preview review token and any required preserve/overwrite decision. The MCP request limit is 1 MiB.

workspace_audience_csv_preview

ChatGPT

Preview column mappings and the reviewed create plan for a CSV audience import. Send unchanged inputs with review_token to apply.

workspace_audience_delete

ChatGPT

Delete a customer-owned static audience and its memberships after confirming its current name. Workspace leads remain.

workspace_audience_get

ChatGPT

Get one customer-owned static audience by ID.

workspace_audience_leads_list

ChatGPT

List active leads in a customer-owned static audience, with bounded pagination.

workspace_audiences_list

ChatGPT

List customer-owned static audiences in a workspace, with bounded pagination.

workspace_billing_overview_get

ChatGPT

Read the customer-visible balance, credits, invoices and billing settings without making a financial commitment.

workspace_billing_platform_notice_get

ChatGPT

Read the customer-visible billing platform notice.

workspace_blocklist_import_apply

ChatGPT

Apply at most 2,000 CSV rows and 1 MiB of encoded arguments in one transaction. Existing active blocks remain unchanged; invalid rows are reported and valid rows can succeed. Requires workspace owner/admin or Lightmeter staff access.

workspace_blocklist_import_preview

ChatGPT

Preview at most 2,000 CSV rows and 1 MiB of encoded arguments without writing. Uses the same header, duplicate, and invalid-row interpretation as Settings. Requires workspace owner/admin or Lightmeter staff access.

workspace_blocklist_values_add

ChatGPT

Add up to 500 email addresses or domains to an active workspace blocklist. Requires workspace owner/admin or Lightmeter staff access. Returns added, duplicate, and invalid outcomes for each value; existing blocks are not duplicated. This does not unsubscribe leads or reopen previously skipped campaign leads.

workspace_blocklist_values_export

ChatGPT

Export a canonical-order page of up to 200 active email or domain blocks as spreadsheet-safe CSV. Every page includes the CSV header and can be imported separately; when combining pages, keep the first header and append only data rows from later pages. Continue with next_cursor and unchanged filters while has_more is true. Reduce limit on result_too_large; if limit 1 still fails, use the public API full CSV stream. Requires workspace owner/admin or Lightmeter staff access.

workspace_blocklist_values_list

ChatGPT

List or search active email-address or domain blocks in one workspace. Requires owner/admin or Lightmeter staff access. Returns at most 200 values per page with total and available sources. On result_too_large reduce limit; if limit 1 still fails, use the public API list or CSV stream.

workspace_blocklist_values_remove

ChatGPT

Remove 1 to 500 active email-address or domain blocks in one workspace. Requires owner/admin or Lightmeter staff access. Returns removed, not_found, duplicate, and invalid outcomes per value. A repeat removal is a harmless not_found result.

workspace_campaign_url_tracking_check

ChatGPT

Request a fresh sending-domain readiness check for campaign click tracking.

workspace_campaign_url_tracking_update

ChatGPT

Enable or disable campaign click tracking for new campaigns or for new and existing campaigns.

workspace_creative_quality_assess

ChatGPT

Return the same advisory copy-quality metrics, guidance, and text findings shown by the Creative editor for submitted subject and body text. This read-only assessment does not save, approve, or send Creative. RANDOM alternatives use expected word counts and worst-alternative spam-trigger scores. Each field is limited to 64 KiB, template nesting to 32 levels, total RANDOM alternatives to 1024, and results to 256 findings; oversized copy is rejected, never truncated.

workspace_crm_activate

ChatGPT

Activate the configured CRM integration. Close requires the current configuration_revision.

workspace_crm_blocklist_approve

ChatGPT

Approve the latest reviewed blocklist synchronization preview. Close requires unchanged settings and configuration_revision.

workspace_crm_blocklist_preview

ChatGPT

Preview blocklist synchronization against the selected CRM. Close requires current settings and configuration_revision.

workspace_crm_capability_activate

ChatGPT

Activate one configured CRM capability. Close requires the current configuration_revision.

workspace_crm_capability_disable

ChatGPT

Disable one configured CRM capability.

workspace_crm_close_get

ChatGPT

Read saved Close connection and sync state without refreshing provider metadata.

workspace_crm_conversation_preview

ChatGPT

Preview the CRM conversation test, optionally for one conversation UUID, without sending the test.

workspace_crm_conversation_retry

ChatGPT

Queue a retry for one CRM conversation UUID.

workspace_crm_conversation_test_get

ChatGPT

Read the durable result of a CRM conversation test by operation UUID.

workspace_crm_conversation_test_start

ChatGPT

Start one durable CRM conversation test. Use the preview values and a new operation_id in payload; query that operation before any retry after an uncertain result.

workspace_crm_disable

ChatGPT

Disable the configured CRM integration and its synchronization capabilities.

workspace_crm_hubspot_get

ChatGPT

Read saved HubSpot connection and sync state without refreshing provider metadata.

workspace_crm_oauth_disconnect

ChatGPT

Disconnect the current OAuth grant. Supply the current oauth_grant_revision; Close also requires configuration_revision.

workspace_crm_revenue_setup_get

ChatGPT

Read the customer-visible Revenue attribution setup and provider metadata for the workspace.

workspace_crm_revenue_setup_update

ChatGPT

With a write grant and Admin or Owner role, save Revenue attribution setup using the public API request body as payload. This may invalidate cached Revenue reports.

workspace_crm_update

ChatGPT

Replace saved CRM mapping settings. Close requires the current configuration_revision.

workspace_crm_validate

ChatGPT

Validate the saved CRM credential and connection against the selected provider.

workspace_domain_generation_update

ChatGPT

Replace the workspace industry keywords used for generated sending-domain suggestions.

workspace_domain_industry_keywords_suggest

ChatGPT

Suggest concise industry keywords from a public customer website for a workspace. Requires workspace settings management permission; fetches the website and uses generation allowance without saving or replacing keywords.

workspace_forwarding_destination_update

ChatGPT

Update and verify the workspace reply-forwarding destination. A new operation_id may send one verification email; reuse after an uncertain result is rejected.

workspace_identity_generation_update

ChatGPT

Replace the workspace market, identity-profile weights, and given-name style used for generated mailboxes.

workspace_identity_samples_generate

ChatGPT

Generate six non-persistent identity samples for proposed market, profile weights, and given-name style.

workspace_infrastructure_campaign_assignments_list

ChatGPT

Traverse active campaign-to-mailbox assignments in bounded pages, including campaign, mailbox, and domain references.

workspace_infrastructure_capacity_get

ChatGPT

Get the workspace's daily capacity, subscribed amount, sent and unused per day over the last 7 days, loaded-lead runway and ramp-up stages, as shown in Infra › Capacity. Set include_days=true for the daily capacity, sent and scheduled series over horizon_days (21 or 42). Lightmeter staff also receive the staff recommendation.

workspace_infrastructure_distribution_list

ChatGPT

List campaign mailbox distribution for the same bounded date range as the Infrastructure web view.

workspace_infrastructure_domain_mail_server_sent_counts_get

ChatGPT

Get workspace-scoped mail-server sent counts for one domain and date range.

workspace_infrastructure_domain_walkin_get

ChatGPT

Read an owned sending domain's destination, walk-in participation, UTM label, readiness and manager capability through the public API contract.

workspace_infrastructure_domain_walkin_update

ChatGPT

Apply a manager-authorized partial domain destination or walk-in setting change through the public API. Read settings first and inspect the returned effective destination and changed flag.

workspace_infrastructure_domains_export

ChatGPT

Export 1 to 500 selected workspace domains as CSV with formula neutralization. Set compact=true when campaign names make a row too large; traverse assignment pages separately.

workspace_infrastructure_domains_list

ChatGPT

List customer-visible domain inventory and health for the same bounded date range as the Infrastructure web view.

workspace_infrastructure_domains_walkin_update

ChatGPT

Enable or disable walk-in participation for 1–200 active workspace-owned domain UUIDs atomically. Requires workspace manager permission; enabling also requires workspace tracking already enabled. Duplicates are applied once. Any validation or ownership failure changes nothing. Affects the entire domain, including unselected mailboxes; preserves destinations and other settings. Saved participation does not imply routing readiness.

workspace_infrastructure_mailbox_send_log_list

ChatGPT

List campaign sends for one workspace mailbox. Warming and staff-only send types are excluded. Page limit is 1 to 100.

workspace_infrastructure_mailboxes_export

ChatGPT

Export 1 to 500 selected workspace mailboxes as CSV without login credentials. Set compact=true when campaign names make a row too large; traverse assignment pages separately.

workspace_infrastructure_mailboxes_list

ChatGPT

List customer-visible mailbox inventory without maintenance credentials.

workspace_infrastructure_onboarding_build_status_get

ChatGPT

Get aggregate domain and mailbox build progress for the selected workspace. Returns no status when no onboarding build exists.

workspace_infrastructure_ramp_up_automation_set

ChatGPT

Lightmeter staff only. Switch 1–500 active workspace mailboxes, selected by up to 500 mailbox_ids or by every mailbox in up to 100 domain_uuids (not both), to Automatic (enabled=true) or Manual (enabled=false) ramp-up management. Automatic adopts each mailbox's current daily limit without reducing it or resetting earned progress; repeating it preserves state. Every limit must be within the captured floor and ceiling, otherwise nothing changes. Manual stops automatic limit management and keeps the current limit. Duplicates are applied once; a selection resolving to more than 500 mailboxes changes nothing.

workspace_infrastructure_ramp_up_domain_mailboxes_list

ChatGPT

Traverse customer-visible ramp-up mailbox rows for one domain in bounded pages; compact mode omits large nested campaign references.

workspace_infrastructure_ramp_up_domains_list

ChatGPT

List customer-visible domain and mailbox health, ramp-up progress, and allocation policy evidence.

workspace_infrastructure_ramp_up_mailbox_get

ChatGPT

Get customer-visible mailbox reputation, volume, projection, and ramp-up history.

workspace_lead_import_undo

ChatGPT

Attempt a conflict-aware undo of a reviewed import. Review import history and set confirm_undo=true. The response reports partial conflicts explicitly.

workspace_lead_imports_list

ChatGPT

List recent lead imports and their reversal status for the workspace.

workspace_lead_reply_history_get

ChatGPT

Get one active workspace lead's campaign and qualified reply history with bounded reply pagination.

workspace_leads_bulk_delete

ChatGPT

Soft-delete 1 to 1,000 selected active workspace leads. Review the exact IDs and set confirm_delete=true. Already deleted IDs are skipped on retry.

workspace_leads_csv_apply

ChatGPT

Apply the unchanged CSV and column mappings using expected_impact from the prior review and a client-generated operation UUID for retry. Explicitly confirm overwrites when required. Campaign assignments in CSV are excluded from this workspace import tool.

workspace_leads_csv_export

ChatGPT

Export up to 100 matching active workspace leads as one CSV page. Continue with next_cursor until has_more is false; each page uses the browser CSV column order and spreadsheet formula escaping.

workspace_leads_csv_impact

ChatGPT

Review exact create, update, unchanged, failed and overwrite counts for the CSV and column mappings before apply.

workspace_leads_csv_preview

ChatGPT

Inspect CSV columns and sample rows without writing. Supply UTF-8 CSV content, at most 10 MiB and 10,000 rows.

workspace_leads_csv_template_get

ChatGPT

Return the browser's CSV template for a workspace lead import.

workspace_leads_import_apply

ChatGPT

Apply one unchanged structured lead import preview only after the whole-source audience inventory is within its limits, every chunk in the complete source has been reviewed, and whole-source identity_refs/resolved-lead components have been checked for cross-chunk overlap. Existing values are preserved unless you explicitly selected overwrite and confirmed the reviewed replacements. Reuse the same operation_id when retrying an unknown outcome.

workspace_leads_import_preview

ChatGPT

Preview creation, additions, protected values, explicit overwrites, and audience assignments for up to 2,000 structured workspace leads in one atomic import. This tool never writes data. For a larger source, require one non-empty unique source_ref per row across the complete source, report and exclude duplicate/missing refs while keeping unrelated valid rows, and split the remainder into deterministic chunks under 2,000 rows and 10 MiB. Before previewing any chunk, inventory all audience definitions and effective assignments across the complete source. Require one identical definition per stable audience ref, at most 100 distinct audience refs, and at most 10,000 unique (source_ref, effective audience ref) assignment edges across the logical import; do not preview any chunk if the complete-source inventory exceeds either limit. After the inventory passes, preview every chunk before applying any chunk. Combine every returned identity_refs value and resolved lead_id into whole-source transitive components; do not apply if one component spans chunks. Complete results are limited to 10 MiB; on arguments_too_large or result_too_large, repartition into smaller deterministic chunks and preview the complete source again.

workspace_leads_list

ChatGPT

Search and filter active workspace leads with cursor pagination. This is the same customer lead list as the browser and public API.

workspace_qualification_cost_estimate

ChatGPT

Read-only qualification triage and estimated costs for supplied companies and a saved template name. This does not start a paid job.

workspace_qualification_job_cancel

ChatGPT

Request cancellation of a pending or running external qualification job.

workspace_qualification_job_export

ChatGPT

Read a base64-encoded CSV chunk. Use next_offset until null to assemble the complete export.

workspace_qualification_job_progress

ChatGPT

Read the current or terminal status of an external qualification job.

workspace_qualification_job_results

ChatGPT

Read a page of stored qualification results when available, including partial results from a failed or cancelled job.

workspace_qualification_job_start

ChatGPT

Start a qualification job after a cost estimate. Provider work consumes existing workspace allowance; cancellation does not refund completed work. Supply a stable operation_id for safe retries and a maximum USD cost ceiling.

workspace_qualification_jobs_list

ChatGPT

List qualification jobs started through the workspace API or MCP.

workspace_qualification_prompt_template_create

ChatGPT

Create a saved qualification prompt template. This does not start a qualification job.

workspace_qualification_prompt_template_default_get

ChatGPT

Get the workspace's default qualification prompt template.

workspace_qualification_prompt_template_default_set

ChatGPT

Set a saved qualification prompt as the workspace default.

workspace_qualification_prompt_template_delete

ChatGPT

Delete a nondefault qualification prompt template.

workspace_qualification_prompt_template_get

ChatGPT

Get one saved qualification prompt template by ID.

workspace_qualification_prompt_template_update

ChatGPT

Replace a saved qualification prompt template by ID.

workspace_qualification_prompt_templates_list

ChatGPT

List the workspace's saved qualification prompt templates.

workspace_reply_forwarding_mode_update

ChatGPT

Choose whether reply forwarding is off, includes all replies, or includes only positive replies.

workspace_reply_forwarding_recipient_add

ChatGPT

Add and check an address without replacing existing recipients. Requires a fresh operation_id for the verification email.

workspace_reply_forwarding_recipient_remove

ChatGPT

Remove only the specified reply forwarding address, preserving other recipients.

workspace_reply_snippets_content_get

ChatGPT

Read a reply snippet's stored text and HTML in bounded UTF-8 chunks. Continue with the returned byte offsets while either has_more flag is true; this also supports older large UI snippets.

workspace_reply_snippets_create

ChatGPT

Create a shared workspace reply snippet. HTML is sanitized by the same rule as the web editor; provide plain text, HTML, or both.

workspace_reply_snippets_delete

ChatGPT

Delete one shared workspace reply snippet by ID. Existing drafts and sent mail are unaffected.

workspace_reply_snippets_list

ChatGPT

List the shared reply snippets in a workspace. Snippet variables are returned as stored; inserting one into a Unibox draft does not send mail.

workspace_reply_snippets_update

ChatGPT

Replace one workspace reply snippet's name and body. This does not change existing drafts or previously sent mail.

workspace_report_campaign_performance_get

ChatGPT

Get the reports campaign performance segmentation read model for a workspace.

workspace_report_campaigns_get

ChatGPT

Get the reports campaigns table read model for a workspace.

workspace_report_deliverability_detail_get

ChatGPT

Get the same detailed deliverability projection as the selected browser report tab. Granularity applies only to boxhost-trends.

workspace_report_deliverability_get

ChatGPT

Get the reports deliverability read model for a workspace and date range.

workspace_report_engagement_performance_get

ChatGPT

Get the reports engagement performance trend read model for a workspace and date range.

workspace_report_engagement_reply_breakdown_get

ChatGPT

Get the reports engagement reply breakdown read model for a workspace and date range.

workspace_report_revenue_availability_get

ChatGPT

Check whether the workspace's verified CRM connection and Revenue settings currently allow the report.

workspace_report_revenue_get

ChatGPT

Read the conditional Revenue report through the same CRM verification, currency, stage, cache, and stale-response contract as the browser.

workspace_report_speed_to_lead_get

ChatGPT

Get the speed-to-lead browser report projection. Defaults to the most recent seven UTC dates; provide both dates for a custom range.

workspace_report_stocks_get

ChatGPT

Get the reports stocks and lead-runway read model for a workspace.

workspace_report_summary_get

ChatGPT

Get the reports summary read model for a workspace and date range.

workspace_report_volume_get

ChatGPT

Get the reports volume read model for a workspace and date range.

workspace_settings_general_get

ChatGPT

Read customer-visible sending and identity-generation settings for a workspace.

workspace_team_member_access_remove

ChatGPT

Remove an existing member's access to the workspace. Requires Admin or Owner.

workspace_team_member_create

ChatGPT

Invite or create a workspace member under the signed-in member's role permissions.

workspace_team_member_role_update

ChatGPT

Replace an existing workspace member's role. Requires Admin or Owner.

workspace_team_member_setup_link_create

ChatGPT

Create a private one-time setup link for an existing workspace member. Requires Admin or Owner.

workspace_team_members_list

ChatGPT

Read the customer-visible team roster and role capabilities for a workspace.

workspace_unibox_access_get

ChatGPT

Lightmeter staff only: read whether customers may use Unibox after workspace deactivation.

workspace_unibox_access_update

ChatGPT

Lightmeter staff only: preserve or revoke customer Unibox access after workspace deactivation. Active workspace access and campaign sending policy stay unchanged.

workspace_walkin_report_get

ChatGPT

Read workspace/domain request totals, estimated bot-filtered navigation, automation, unknown and non-navigation counts. Null estimates mean unavailable measurement, not zero humans. Current-campaign associations do not prove the campaign that caused a visit or identify a recipient. Results are limited to 128 KiB; narrow dates/domain or use the public API for a larger report.

workspace_walkin_tracking_update

ChatGPT

Update workspace walk-in visitor tracking, optional UTM-source capture, or the default redirect destination. Supply at least one setting.

workspace_webhook_create

ChatGPT

Create a workspace webhook for one or more event types. A signed format returns its signing secret once.

workspace_webhook_delete

ChatGPT

Delete one workspace-owned webhook configuration.

workspace_webhook_deliveries_list

ChatGPT

Read paged webhook delivery history for a workspace.

workspace_webhook_delivery_get

ChatGPT

Read one workspace-owned webhook delivery and its attempts.

workspace_webhook_preview_test_send

ChatGPT

Send sample events to an unsaved public HTTPS destination without recording delivery history. Use a unique operation_id and inspect the receiver after an uncertain result.

workspace_webhook_signing_secret_rotate

ChatGPT

Rotate a signed webhook's secret and return the new secret once.

workspace_webhook_test_send

ChatGPT

Send sample events to a saved webhook and record delivery history. Use a unique operation_id; after an uncertain result, inspect delivery history before another send.

workspace_webhooks_list

ChatGPT

List a workspace's webhook configurations without returning signing secrets.

workspaces_capacity_planning_list

ChatGPT

Lightmeter staff only: list every active workspace with daily capacity, sent per day, subscribed amount, loaded-lead runway, spare mailboxes, share held for health and needs, sorted by urgency, as on the staff Planning page.