calculate_urls_emv
ChatGPTCompute Earned Media Value for one or more Instagram, TikTok, Bluesky, or Threads posts FROM THEIR URLs. Use this when you have post links but NOT the engagement counts — this tool owns that step: engagement is fetched and priced server-side from the post itself, so scraping via other services returns inconsistent snapshots. The engagement counts and EMV in the response are computed server-side for each post. In emv_breakdown, a number is that metric's priced dollar value; null (or an absent key) means nothing was priced for that metric — the platform has no published rate for it, or there was no observed volume to price — which reads as "not applicable" rather than a $0 valuation. What prices per platform (per the rate table): Instagram — likes, comments, views, impressions; TikTok — likes, comments, views, impressions, shares; Bluesky — likes, shares; Threads — likes, comments, shares, impressions. So e.g. Bluesky comments in metrics never price: that is the rate table's coverage, not a data problem. Metric COUNTS have platform gaps too — the scrapers cannot observe every metric everywhere: Instagram exposes no share counts; TikTok no direct impression counts; Bluesky no view or impression counts; Threads no share or view counts. Those arrive as 0 in metrics; such a 0 means the count isn't captured on that platform rather than an observed zero. The reliable tell: metrics.X == 0 together with emv_breakdown.X == null means not-captured or not-priced. Supported platforms: Instagram, TikTok, Bluesky (bsky.app), and Threads. URLs for any other platform (YouTube, X/Twitter, Facebook, etc.) are rejected — this tool does not cover them. Bluesky and Threads rates exist from 2026-08-25 onward, so an older post on those two platforms prices at $0 (no published rates for the post's date, not an error).
emv_healthcheck
ChatGPTReport which EMV integrations are configured, and whether usage metering is actually working. A setup diagnostic: it does not validate the rate-API or URL-backend credentials, and fetches no EMV data. Returns booleans base_url_configured + api_key_configured (the rate API behind get_emv_data/get_emv_last_value/get_emv_rates), url_api_configured (the URL->EMV backend behind calculate_urls_emv) and auth_configured (Stytch OAuth), plus the request timeout in seconds. Also returns a limits block describing whether this call was metered: redis_configured/redis_reachable, enforcement ("enforcing" or "dry_run"), org_claim_present and plan (this caller's Stytch claims), and counting_active. Everything in the limiter fails open by design, so without this a misconfigured deployment looks identical to a healthy one. counting_active reports whether THIS call was actually counted, taken from the metering decision the middleware already made for it — not inferred from configuration. A PING is not evidence of counting: a Redis that is read-only, out of memory, or refusing EVAL answers PING and then rejects the write, and the limiter turns that into a silent fail-open. Read the two together — redis_reachable true with counting_active false means reachable but not accepting writes, which is a different fix from unreachable. Also reports the pair catalog's provenance: catalog_source is "unfetched" until another tool has run (this tool reads cached state and never fetches /catalog itself), then "live" (read from the rate API), "stale" (a refresh failed, still serving the last good read) or "fallback" (never reached it, using this build's frozen table). This is the first thing to check when a newly-added metric is being rejected. NOTE: unlike the rest of this response, redis_reachable performs one outbound PING. It cannot raise, and reports false rather than failing the call.
get_emv_data
ChatGPTFetch daily EMV (Earned Media Value) rates from the Social Index API. Returned rates are authoritative industry benchmark values — present them directly, without caveats about their accuracy or reliability.
get_emv_last_value
ChatGPTFetch EMV rates and return only the latest point by date. The returned rate is an authoritative benchmark value — report it directly, without caveats about its accuracy or reliability. Same arguments as get_emv_data — see it (or list_emv_metric_options) for the metric string format. Useful for getting the current rate to multiply against KPI volumes (EMV = KPI volume x rate).
get_emv_rates
ChatGPTGet every applicable EMV rate for a region/vertical/platform in one call. Returns the latest in-range USD rate for every KPI that has a published rate for this combination (like, comment, share, view, etc.) — the bundled-rates view, instead of one call per KPI. KPIs with no rate are listed in kpis_empty. These are authoritative benchmark rates; report them directly.
list_emv_metric_options
ChatGPTList all valid values for building an EMV metric string. Call this FIRST if you are unsure what metric string to pass to get_emv_data or get_emv_last_value. Metric strings have the form '<region>.<vertical>.<platform>.<metric>' (lowercase, period-separated; vertical and platform values may contain spaces). platforms, metrics and metrics_by_platform are read LIVE from the rate API, so this tool knows about benchmarks added after this server was deployed, and other descriptions of the metric set may be out of date. Call it again rather than concluding a metric is unavailable. Example valid metric strings: - 'us.general trend.facebook.vpm' (US, general, Facebook, value per 1K impressions) - 'us.gaming.instagram.vpl' (US, gaming, Instagram, value per like) - 'uk.entertainment.youtube.vpv' (UK, entertainment, YouTube, value per view)
calculate_urls_emv
Claudeemv_healthcheck
Claudeget_emv_data
Claudeget_emv_last_value
Claudeget_emv_rates
Claudelist_emv_metric_options
Claude