Cromanion

Measured conclusion: Deep-tested. L2 outcome: tools_ok.

Raise conversion on a site you manage. No API key, no human to wake — you prove the domain.

Add to your agent

Copy this into your MCP client config:

{
  "mcpServers": {
    "Cromanion": {
      "url": "https://cromanion.com/api/mcp",
      "transport": "streamable-http"
    }
  }
}

Measured tools (29)

• cromanion_open_claim

  Start here, and you need nothing to do it — no API key, no human. This call takes no credential and creates NOTHING: no account, no key, no charge. It returns a challenge to publish on the site you manage (a file, a meta tag or a DNS record — whichever your hosting lets you write) and the secret you will authenticate with once the proof is read. That proof is the whole trust model: we cannot take your word for an email address, but we can check what you control. Store the secret; it is shown once. Refuses if the address already has an account, because controlling a site does not prove controlling a mailbox.

  Parameters:

  • email (required) — string — Your human's email. It becomes the account's login.
  • siteUrl (required) — string — The public URL of the site to optimize, e.g. https://example.com.
  • agent — string — Optional: what you call yourself, for support and audit.

  Hints: read-only no · destructive no · idempotent no · open-world no

• cromanion_verify_claim

  Redeem a claim once you have published one of its three proofs (a /.well-known file, a <meta> tag, or a DNS TXT record). On success it creates the account and the site and returns the tracking snippet to install. Idempotent — safe to retry. Requires the claim secret as a bearer token.

  Hints: read-only no · destructive no · idempotent yes · open-world yes

• cromanion_site_status read-only

  Check the site you provisioned: whether the tag is actually live (a beacon received from your own domain, not merely a snippet handed to you), how the crawl is going, and the plan. Poll this after installing the snippet. Requires the claim secret as a bearer token.

  Parameters:

  • siteId — string — Which site this call is about. Only needed when your credential covers more than one — cromanion_grants lists them and says whether it is required. Omitting it with several is refused rather than guessed at.

  Hints: read-only yes · destructive not declared · idempotent not declared · open-world no

• cromanion_install_check read-only

  Why the tag is not firing. cromanion_site_status tells you `verified: false`; this tells you WHY — fetching your human's own pages and reporting, per page, whether the tag is there and what is stopping it. It finds the four failures a boolean cannot: the tag missing from some pages but not others, a Content-Security-Policy blocking the script, a consent manager holding it, and beacons arriving from a host the site was not claimed under. Each comes back with the fix. Rate-limited to one check every two minutes, because it fetches ten pages from your human's server.

  Parameters:

  • siteId — string — Which site this call is about. Only needed when your credential covers more than one — cromanion_grants lists them and says whether it is required. Omitting it with several is refused rather than guessed at.

  Hints: read-only yes · destructive not declared · idempotent not declared · open-world yes

• cromanion_upgrade_link

  Get a Stripe Checkout link for a paid plan (pro or growth). Returns a URL and nothing else — YOU do not pay and cannot: give the link to your human, who enters their own card. The site works on the free plan meanwhile. Requires the claim secret as a bearer token.

  Parameters:

  • plan (required) — string — Which paid plan the link should check out: pro or growth. The site keeps working on free until your human completes it.
  • siteId — string — Which site this call is about. Only needed when your credential covers more than one — cromanion_grants lists them and says whether it is required. Omitting it with several is refused rather than guessed at.

  Hints: read-only no · destructive no · idempotent no · open-world yes

• cromanion_errors read-only

  Every refusal this server can return, and what to do about each. Needs no credential — you can read it before you have one, or after a refusal has told you the one you had is gone. The field to branch on is `retry`: `later` means a cooldown or a transient, `fix_request` means change something and resend, `needs_human` means nothing you send will help until a person acts. Read it once and stop guessing whether a 409 is worth retrying.

  Hints: read-only yes · destructive not declared · idempotent not declared · open-world no

• cromanion_grants read-only

  List what your human has allowed this agent to do, and what they have not. Read this before assuming a capability — the alternative is discovering your permissions by failing in front of them. You cannot change these yourself; only your human can, in their console.

  Hints: read-only yes · destructive not declared · idempotent not declared · open-world no

• cromanion_pending read-only

  Everything on this site that is waiting for a decision: page rewrites the agent wants to make, rules it distilled from real sessions, conversions it spotted but has never counted, and pages your traffic found that the crawl missed. This is the client's actual job in the console, in one call. Rule on an item with cromanion_decide.

  Parameters:

  • siteId — string — Which site this call is about. Only needed when your credential covers more than one — cromanion_grants lists them and says whether it is required. Omitting it with several is refused rather than guessed at.

  Hints: read-only yes · destructive not declared · idempotent not declared · open-world no

• cromanion_config read-only

  Everything IN FORCE on this site: the conversions that count, the facts you may state and the ones forbidden, which surfaces are enabled, the brand, and every distilled rule WITH its status. cromanion_pending shows what is waiting for a decision; this shows what was already decided — including by you. Read it before writing anything, because most writes here REPLACE rather than add.

  Parameters:

  • siteId — string — Which site this call is about. Only needed when your credential covers more than one — cromanion_grants lists them and says whether it is required. Omitting it with several is refused rather than guessed at.

  Hints: read-only yes · destructive not declared · idempotent not declared · open-world no

• cromanion_decide

  Rule on one thing from cromanion_pending: pass its `kind` and `id` back unchanged with a decision. `approve` / `reject` for everything; `pause` additionally for a playbook rule. Approving a GOAL is different in kind from the rest — it changes what every impact figure for this site means, permanently. You cannot edit rule text here, only rule on it.

  Parameters:

  • kind (required) — string — Which queue the item came from. Copy it back from cromanion_pending unchanged.
  • decision (required) — string — A status and nothing else — rule TEXT is edited by cromanion_set_playbook. `pause` applies only to a playbook_rule.
  • id — string — The item's id, exactly as cromanion_pending returned it. Ids are not shared across kinds.
  • all — boolean — Rule on every discovered page at once, ignoring `id`. observed_path only — the other kinds are decided one at a time, as in the console.
  • siteId — string — Which site this call is about. Only needed when your credential covers more than one — cromanion_grants lists them and says whether it is required. Omitting it with several is refused rather than guessed at.

  Hints: read-only no · destructive no · idempotent yes · open-world no

• cromanion_set_playbook destructive

  Reword a distilled rule, or ask for a fresh distillation. Rules are what the agent learned from real sessions; cromanion_decide approves or pauses them, and this changes what they SAY. Send a ruleId with condition and/or recommendedAction to reword one — ids come from cromanion_config, which also shows rules you have already approved or paused. Send redistill: true instead to distil new rules from recent sessions; they arrive as `proposed` for you to rule on. Rate-limited to one distillation an hour. Needs the decisions permission.

  Parameters:

  • ruleId — string — The rule to reword, from cromanion_config.
  • siteId — string — Which site this call is about. Only needed when your credential covers more than one — cromanion_grants lists them and says whether it is required. Omitting it with several is refused rather than guessed at.
  • condition — string — When the rule applies, in plain words. Omit to leave it.
  • redistill — boolean — Ignore ruleId and distil fresh rules from recent sessions instead.
  • recommendedAction — string — What the agent should do when it applies. Omit to leave it.

  Hints: read-only no · destructive yes · idempotent no · open-world no

• cromanion_ask_human

  Ask your human for something only they can do, in a place they will actually see it. Every other refusal here ends with YOU reading it — a permission they did not grant, a payment you cannot make, a mode you may not set — and from their side you simply went quiet. This puts one request on their notifications screen, attributed to you. Nothing replies here: they act in their console, and you find out by reading cromanion_grants or cromanion_config again. One request an hour, and a second from you replaces the first rather than stacking.

  Parameters:

  • title (required) — string — What you need, in a line they can read at a glance. Up to 90 characters.
  • body (required) — string — Why you need it and what it unblocks, in a sentence or two. Up to 400 characters.
  • siteId — string — Which site this call is about. Only needed when your credential covers more than one — cromanion_grants lists them and says whether it is required. Omitting it with several is refused rather than guessed at.

  Hints: read-only no · destructive no · idempotent yes · open-world no

• cromanion_mark_read

  Mark notifications read once you have passed them on to your human. cromanion_pending returns the unread ones and nothing cleared them, so the queue grew for ever and you re-read the same items on every poll. Send a notificationId for one, or nothing to clear them all. Needs the decisions permission.

  Parameters:

  • siteId — string — Which site this call is about. Only needed when your credential covers more than one — cromanion_grants lists them and says whether it is required. Omitting it with several is refused rather than guessed at.
  • notificationId — string — One notification, by the id cromanion_pending returned. Omit to mark every unread one read.

  Hints: read-only no · destructive no · idempotent yes · open-world no

• cromanion_impact read-only

  What the agent has actually changed: visitors who saw it versus the 10% held back as a control. Read `liftStatus` before reporting anything — `learn` means nothing has been shown yet and `insufficient` means the arms are too small to tell signal from noise. Neither is 'no lift'.

  Parameters:

  • siteId — string — Which site this call is about. Only needed when your credential covers more than one — cromanion_grants lists them and says whether it is required. Omitting it with several is refused rather than guessed at.

  Hints: read-only yes · destructive not declared · idempotent not declared · open-world no

• cromanion_visits read-only

  List your human's recent visits — each with its pages, how it ended, what Cromanion's agent showed or held back, the doubts it read, and the start of the visit's story. Filter by page, outcome, doubt, device, channel and more; page through with `next`. Needs the visits permission, which only your human can give in their console. Everything under `reading` is Cromanion's interpretation, not something the visitor said — report it that way. A visit is readable in full about 30 minutes after the visitor leaves; for who is on the site right now use cromanion_live. The control group is never read, so a doubt filter contains no control visits. Shares a budget of 60 calls an hour with cromanion_visit; every call is on your human's record before it runs. Identifiers in paths are masked (`:id`, `:token`), so products addressed by a number or a code are indistinguishable to a page filter.

  Parameters:

  • page — string — Visits that viewed this path — exact, or a prefix ending in `*` (`/products/*`). Pages your human excluded are refused, not answered with zero.
  • limit — number — 1 to 50. Default 20. Newest first.
  • since — string — ISO date. Visits that STARTED at or after it. Default: 24 hours ago. At most 30 days back.
  • until — string — ISO date. Visits that started before it. Default: now.
  • cursor — string — The `next` of the previous page.
  • device — string — The device the visit arrived on.
  • siteId — string — Which site this call is about. Only needed when your credential covers more than one — cromanion_grants lists them and says whether it is required. Omitting it with several is refused rather than guessed at.
  • channel — string — How the visit was acquired — a closed label, never a campaign name.
  • outcome — string — `left` = ended without your human's conversion. Default `any`.
  • visitor — string — First visit, or a returning visitor (a coarse, cookieless reading).
  • objection — string — Visits where the model read this doubt at least once. A reading, not a fact.
  • population — string — People (default), AI shopping agents, or both. Crawlers and scripts are never listed.
  • agentAction — string — Visits where the agent's intervention was seen, sent, held back, or where the site was still learning. Default `any`.

  Hints: read-only yes · destructive not declared · idempotent not declared · open-world no

• cromanion_visit read-only

  One visit in full: the story of each page load, every moment Cromanion's agent looked at the visitor and what it decided, what the visitor actually saw, and the doubts it read. Pass the `visitId` from cromanion_visits, or a `sessionId` from cromanion_live once that visitor has left. Needs the visits permission. Readings sit under `reading`; a moment at or after the conversion carries `observed` instead, because the conversion is a fact. `shown.seen` is what the visitor's browser confirmed, not what was sent. Same 60-an-hour budget as cromanion_visits.

  Parameters:

  • siteId — string — Which site this call is about. Only needed when your credential covers more than one — cromanion_grants lists them and says whether it is required. Omitting it with several is refused rather than guessed at.
  • visitId — string — From cromanion_visits.
  • sessionId — string — Instead of visitId: any session of the visit, e.g. one seen in cromanion_live.

  Hints: read-only yes · destructive not declared · idempotent not declared · open-world no

• cromanion_live read-only

  Who is on your human's site right now: each visitor's pages so far, what the agent showed them, and its current reading of them. Paid plans only (Pro and up) — on Free it is refused with the way to upgrade. Needs the visits permission. Poll at most once a minute; this has its own budget, separate from the history tools. Bots are judged when a visit ends, so a robot not yet recognised can appear here as a person.

  Parameters:

  • siteId — string — Which site this call is about. Only needed when your credential covers more than one — cromanion_grants lists them and says whether it is required. Omitting it with several is refused rather than guessed at.
  • population — string — People (default), AI shopping agents, or both.

  Hints: read-only yes · destructive not declared · idempotent not declared · open-world no

• cromanion_set_mode

  Switch the agent on. `learn` watches and shows visitors nothing; `act` lets it intervene in your human's name. `customizeMode` is a separate axis for rewriting their own page copy: `off`, `shadow` (propose only) or `live`. Needs the activation permission.

  Parameters:

  • mode — string — `learn` watches and shows visitors nothing; `act` lets the agent intervene in your human's name. Omit to leave it as it is.
  • siteId — string — Which site this call is about. Only needed when your credential covers more than one — cromanion_grants lists them and says whether it is required. Omitting it with several is refused rather than guessed at.
  • customizeMode — string — Rewriting your human's own page copy, a separate axis from `mode`: `off`, `shadow` (propose only, nothing shown) or `live`. Omit to leave it as it is.

  Hints: read-only no · destructive no · idempotent yes · open-world no

• cromanion_add_goal

  Define a conversion to count, or change one you already defined. This is not one more setting: it is the numerator of every impact figure your human will read, so a goal that fires on every pageview makes the agent look brilliant and tells them nothing. Send `type` and `value` to create — idempotent, the same matcher twice creates one goal. Send `goalId` instead to rename, reprioritise or price an existing one, or with `remove: true` to stop counting it. Needs the goals permission.

  Parameters:

  • type — string — Matcher: css_click, form_submit, page_view, mailto, tel, outbound, scroll_depth, dwell, custom_event.
  • label — string — What your human will call it. Defaults to something readable.
  • value — string — What the matcher matches — a selector, a path, a percentage, seconds.
  • goalId — string — Change an EXISTING goal instead of creating one. Ids come from cromanion_config. With `remove: true` it stops counting that conversion; otherwise it applies the label, priority or dollarValue you send.
  • remove — boolean — With goalId: stop counting that conversion. Nothing already measured is rewritten.
  • siteId — string — Which site this call is about. Only needed when your credential covers more than one — cromanion_grants lists them and says whether it is required. Omitting it with several is refused rather than guessed at.
  • priority — number — With goalId: 1 is the primary goal, ascending. Rejected rather than clamped if out of range.
  • dollarValue — string — Optional worth of one conversion, for revenue figures.

  Hints: read-only no · destructive no · idempotent yes · open-world no

• cromanion_crawl

  Re-read the site, and name the pages that matter most so they are read first. You know this site's structure better than its owner does — seeding it is the cheapest real improvement to what the agent understands. Rate-limited: one crawl per 30 minutes. Needs the site-config permission.

  Parameters:

  • siteId — string — Which site this call is about. Only needed when your credential covers more than one — cromanion_grants lists them and says whether it is required. Omitting it with several is refused rather than guessed at.
  • priorityUrls — array — Up to 10 paths or URLs to read first, e.g. ["/pricing", "/checkout"]. Anything unusable is handed back rather than dropped.

  Hints: read-only no · destructive no · idempotent no · open-world yes

• cromanion_add_site

  Add another of your human's sites to their account, on a plan that allows more than one. Reaching a site they already own needs no proof — they added it. CREATING one does: nothing about holding a credential for one domain says anything about another, so publish this claim's challenge on the new domain first, the same three ways that proved the first. Refused if your human restricted this credential to specific sites: a list they set should not grow by your action — use cromanion_ask_human. Needs the site-config permission.

  Parameters:

  • siteUrl (required) — string — The public https URL of the site to add, e.g. https://second-shop.com.
  • siteId — string — Which site this call is about. Only needed when your credential covers more than one — cromanion_grants lists them and says whether it is required. Omitting it with several is refused rather than guessed at.

  Hints: read-only no · destructive no · idempotent no · open-world yes

• cromanion_claim_domain destructive

  Claim another host for this site so its traffic counts, or give one up with `remove: true`. A DIFFERENT registrable domain must be proven on its own — publish this claim's challenge there too, then call again. Subdomains need no call at all: a claimed domain already covers them. Needs the site-config permission.

  Parameters:

  • domain (required) — string — A public hostname, e.g. shop.example.com.
  • remove — boolean — Give the host up instead. Needs no proof — it only narrows what we listen to.
  • siteId — string — Which site this call is about. Only needed when your credential covers more than one — cromanion_grants lists them and says whether it is required. Omitting it with several is refused rather than guessed at.

  Hints: read-only no · destructive yes · idempotent yes · open-world yes

• cromanion_set_voice destructive

  Write the facts the agent may state to visitors, and the things it must never say. Plain sentences, one idea per line; start a line with 'never' to forbid something. Anything you write that your human's own pages do not already state is attributed to THEM as the source — so write only what they have told you is true. Every rule is checked against your own text and the site's vocabulary, and anything ungrounded is silently dropped: read the returned `lines` to see what actually survived. Your human sees each line attributed to you. Needs the voice permission.

  Parameters:

  • text (required) — string — The whole guidance, replacing what is there. An empty string clears it.
  • siteId — string — Which site this call is about. Only needed when your credential covers more than one — cromanion_grants lists them and says whether it is required. Omitting it with several is refused rather than guessed at.

  Hints: read-only no · destructive yes · idempotent yes · open-world no

• cromanion_set_surfaces destructive

  Choose where the agent may speak and what those places look like: which layout surfaces are enabled, whether the exit modal is allowed, custom CSS, and per-surface HTML. Markup and CSS are re-checked against an allowlist on every visitor — anything that could run, or fetch a remote resource, is silently removed, so compare what comes back with what you sent. Needs the surfaces permission.

  Parameters:

  • siteId — string — Which site this call is about. Only needed when your credential covers more than one — cromanion_grants lists them and says whether it is required. Omitting it with several is refused rather than guessed at.
  • enabled — array — Surfaces the agent may use: inline_hint, card, sticky_bar, bottom_sheet, inline_block, cta_tooltip.
  • customCss — string — CSS applied inside the intervention's shadow root. An empty string clears it.
  • templates — object — Per-surface HTML, keyed by surface. Use {{message}}, {{cta_label}} and {{cta_url}} as placeholders.
  • allowModal — boolean — Allow the page-blocking exit modal. Off by default.

  Hints: read-only no · destructive yes · idempotent yes · open-world no

• cromanion_set_holdout destructive

  Set the share of visitors held back from the agent so its effect stays measurable (0.05–0.5). Changing it splits the measurement window: periods either side are compared separately and never pooled. Needs the measurement permission.

  Parameters:

  • share (required) — number — Fraction held back, e.g. 0.1 for 10%.
  • siteId — string — Which site this call is about. Only needed when your credential covers more than one — cromanion_grants lists them and says whether it is required. Omitting it with several is refused rather than guessed at.

  Hints: read-only no · destructive yes · idempotent yes · open-world no

• cromanion_set_exclusions destructive

  Put an excluded device back into the measurement. Your human can mark their own devices as "not a real visitor" so their browsing does not count; this removes one of those marks, by the label they gave it. You cannot ADD an exclusion — that names the caller's own browser, which you do not have. Needs the measurement permission.

  Parameters:

  • label (required) — string — The label your human gave the device, as cromanion_config lists it. An ambiguous label is refused rather than guessed at.
  • siteId — string — Which site this call is about. Only needed when your credential covers more than one — cromanion_grants lists them and says whether it is required. Omitting it with several is refused rather than guessed at.

  Hints: read-only no · destructive yes · idempotent yes · open-world no

• cromanion_set_brand

  Adjust the palette, fonts and button styling the agent paints its surfaces with. The crawl guesses these from a stylesheet; you probably know the real tokens. Cosmetic only — it changes how the agent looks, never what it may say. Needs the brand permission.

  Parameters:

  • colors — object — Hex per role, e.g. {"primary":"#1a1a2e"}. Roles come from cromanion_site_status.
  • siteId — string — Which site this call is about. Only needed when your credential covers more than one — cromanion_grants lists them and says whether it is required. Omitting it with several is refused rather than guessed at.
  • ctaFill — string — Primary button background, hex.
  • ctaText — string — Primary button text colour, hex.
  • bodyFont — string — CSS font-family stack for body text, e.g. "Inter, sans-serif".
  • ctaRadius — string — Primary button corner radius, e.g. 8px.
  • headingFont — string — CSS font-family stack for headings, e.g. "Inter, sans-serif".

  Hints: read-only no · destructive no · idempotent yes · open-world no

• cromanion_set_lead_destination destructive

  Set where captured leads are delivered: an https webhook, a connected HubSpot portal, or off. Note before you use it: this is the one change that leaves no other trace — forms keep working and visitors keep being thanked while the addresses go somewhere new — so every change emails your human and appears on their notifications screen. Needs the integrations permission.

  Parameters:

  • kind (required) — string — Where leads go: `webhook` posts them to your `target` URL, `hubspot` to the connected portal, `off` stops delivery while forms keep working.
  • siteId — string — Which site this call is about. Only needed when your credential covers more than one — cromanion_grants lists them and says whether it is required. Omitting it with several is refused rather than guessed at.
  • target — string — The https webhook URL. Not needed for hubspot or off.

  Hints: read-only no · destructive yes · idempotent yes · open-world yes

• cromanion_disconnect_hubspot destructive

  Sever the account's HubSpot connection. The other half of cromanion_set_lead_destination's `hubspot` option: you can point leads at a CRM, so you can take them back out. Refused while any site on the account still delivers there — change those destinations first, or the forms would keep working while every lead failed delivery. Reconnecting is NOT a tool: it is an OAuth consent screen only your human can click, so treat this as one-way and say so before you call it. Needs the integrations permission.

  Parameters:

  • siteId — string — Which site this call is about. Only needed when your credential covers more than one — cromanion_grants lists them and says whether it is required. Omitting it with several is refused rather than guessed at.

  Hints: read-only no · destructive yes · idempotent yes · open-world yes

Publisher website

Measured by MCPYP probing. Untested fields are omitted, never estimated. Listing is always free; MCPYP never proxies calls or stores user tokens.