Peec AI MCP

Archive up to 50 prompts (sets is_archived = true) so they stop running while keeping their chats and history intact. Prefer this over delete_prompts when the data should be retained. Returns a count of archived prompts plus any skipped (not found) or rejected items. Confirm with the user before calling.

Untrusted display-only summary

Assign a built-in classification (Corporate, Competitor, Editorial, Institutional, Other, Reference, UGC, You, Related) or the name of a custom domain classification to a domain. Overrides any heuristic classification. The override applies to the whole root domain, including its subdomains. If the name you pass matche…

Untrusted display-only summary

Assign a built-in classification (Homepage, Category Page, Product Page, Listicle, Comparison, Profile, Alternative, Discussion, How-To Guide, Article, Other) or the name of a custom URL classification to a URL. Overrides any heuristic classification. The override applies to the URL in its normalized form, so it match…

Untrusted display-only summary

Create up to 50 brands (competitors or own) in a project in one call. Returns the created brands with their new ids, plus any skipped items. After creation, call list_brands to confirm the requested brands exist before reporting success, even when they appear in this tool's created list. Duplicates are matched case-in…

Untrusted display-only summary

Create product categories. Each item needs a name and an optional parent_id — omit it for a top-level category, or pass another category's id (use list_categories to find it) to nest it underneath. A name must be unique among its siblings. Items apply in order. Returns per-item results { created, rejected } — the batc…

Untrusted display-only summary

Define a new custom domain classification on a project. This creates the classification entity — it does not assign it to any domain (use assign_domain_classification for that). Confirm with the user before calling. Returns the created classification as { name, color }.

Untrusted display-only summary

Create a brand and return its global_brand_id. Use this when list_global_brands finds no match for a brand you need, or when create_products rejects an item with brand_not_found. Returns { id, name }. The returned name is the brand's canonical name and may differ slightly from the name you sent. This is separate from…

Untrusted display-only summary

Create catalog products (source=CATALOG) in a project — the brand's own claimed/uploaded products, as opposed to LLM products that Peec auto-detects in shopping answers. Each item needs a global_brand_id (use list_global_brands to find it by name, or create_global_brand if the brand doesn't exist yet) and a name uniqu…

Untrusted display-only summary

Create up to 50 prompts in a project in one call. Returns the created prompts with their new ids, plus any skipped/rejected items. Every prompt needs a topic: pick the best-fitting existing topic, or create one with create_topics first. Accepts existing topic_id and tag_ids only — this tool does not auto-create topics…

Untrusted display-only summary

Create up to 50 tags in a project in one call. Each tag may specify a group (grouped tags share the group's color, so their color field is ignored). Returns the created tags with their new ids, plus any skipped items. Duplicates are matched case-insensitively on name.

Untrusted display-only summary

Create up to 50 topics in a project in one call. Topics group related prompts. Returns the created topics with their new ids, plus any skipped/rejected items. Duplicates are matched case-insensitively on name. Items beyond the project's topic limit land in `rejected`.

Untrusted display-only summary

Define a new custom URL classification on a project. This creates the classification entity — it does not assign it to any URL (use assign_url_classification for that). Returns the created classification's name, which is what you pass to assign_url_classification. Confirm with the user before calling.

Untrusted display-only summary

Delete up to 50 brands in a project. Returns a count of deleted brands plus any skipped items (successful items are not echoed back). This is destructive — always confirm with the user before calling.

Untrusted display-only summary

Delete categories by id (use list_categories to resolve ids). A deleted category's child categories and products move up to its parent (a top-level delete sends children to the top level and its products to Uncategorized). Returns per-item results { deleted, rejected } — the batch never fails as a whole. Rejection rea…

Untrusted display-only summary

Permanently delete a custom domain classification entity. Cascades through the override table — any domains currently assigned this classification fall back to their heuristic classification. To only clear a single domain's assignment, use unassign_domain_classification instead. This is destructive — always confirm wi…

Untrusted display-only summary

Delete products by id (use list_products to resolve ids). Returns per-item results { deleted, skipped } — ids not present in the project are skipped (not_found). Only delete source=CATALOG products. Deleting LLM-source products is futile — the ingestion pipeline re-creates an LLM product whenever a shopping answer men…

Untrusted display-only summary

Delete up to 50 prompts in a project. Deletions run asynchronously — returns a count of queued deletions plus any skipped (not found / already deleted) or rejected items (queued items are not echoed back). This is destructive — always confirm with the user before calling.

Untrusted display-only summary

Delete up to 50 user-defined tag groups. By default each group's tags are kept and simply ungrouped; set delete_tags=true on a group to delete its tags and remove them from every prompt. Returns each group with its affected tag count, plus any skipped (not found) or rejected items. This can be destructive — confirm wi…

Untrusted display-only summary

Delete up to 50 tags in a project. Removes tag associations from prompts. Returns a count of deleted tags plus any skipped items (successful items are not echoed back). This is destructive — always confirm with the user before calling.

Untrusted display-only summary

Delete up to 50 topics in a project. Detaches associated prompts (prompts are kept) and deletes any prompt suggestions linked to the topics. Returns a count of deleted topics plus any skipped items (successful items are not echoed back). This is destructive — always confirm with the user before calling.

Untrusted display-only summary

Permanently delete a custom URL classification entity. Any URLs currently assigned this classification fall back to their heuristic (default) classification. To only clear a single URL's assignment, use unassign_url_classification instead. This is destructive — always confirm with the user before calling. Returns { su…

Untrusted display-only summary

Rename, delete and reassign brand-perception attribute clusters — the same edits the Manage attributes screen makes. Resolve every cluster_id and attribute with list_brand_perception_attribute_clusters first; both must be copied verbatim. Pass any combination of renames, delete_cluster_ids and reassignments; at least…

Untrusted display-only summary

Resolve prompt suggestions listed by list_prompt_suggestions, in one pass over a reviewed set: ids in accepted_ids become tracked prompts and promote their topics into the project, ids in rejected_ids are deleted from the review list. Pass either or both, up to 50 ids in total. Rejecting never touches tracked prompts,…

Untrusted display-only summary

Generate prompt suggestions for a project — the same generation Prompt discovery runs in the app. Generation runs in the background: the call returns {queued: true, generation_id}, which get_prompt_suggestion_generation reads until the run succeeds or fails. List successful results with list_prompt_suggestions, then a…

Untrusted display-only summary

Compose candidate topics for the project from its brand profile — the same composition Prompt discovery runs in the app. Nothing is saved: review the returned names with the user and create the chosen ones with create_topics. Composition already accounts for the project's existing topics, so it proposes only what fill…

Untrusted display-only summary

Get Peec's opportunity-scored action recommendations for improving brand visibility in AI search engines. **Always call with `scope=overview` first** to see which slices have the biggest opportunity, then drill down into `owned`, `editorial`, `reference`, or `ugc` with the surfaced url_classification or domain. ## Req…

Untrusted display-only summary

Aggregate AI agent visit counts from access logs over a date range. Without group_by: returns total visit count as a single {visits: N} row. With group_by: returns one row per distinct value of the chosen dimension(s), sorted by visits descending. group_by dimensions: - bot_id: break down by bot (use list_bots to reso…

Untrusted display-only summary

How the project's own brand ranks against competitors for each brand-perception attribute. ranking is the own brand's position among every brand scored for the attribute, ordered by average prominence across all AI answers about it — lower is better, #1 means no brand is more prominent; answers the brand never surface…

Untrusted display-only summary

The sources feeding one brand-perception attribute: the URLs AI answers cited when describing the brand with this attribute. Pass the attribute name exactly as returned by get_brand_perception_brand_attributes or get_brand_perception_attribute_rankings; an unknown name returns an empty result. Data is aggregated acros…

Untrusted display-only summary

How AI models describe the project's own brand: attribute clusters (e.g. "Luxury", "Racing Heritage") with a prominence score. score is 0-100 — the attribute's average prominence per AI answer; higher means the attribute is mentioned earlier and more often. Display scores as-is; they are NOT ratios, never multiply by…

Untrusted display-only summary