Emailchaser

Removing the models leaves a complete and saleable product. Unlimited sending accounts, inbox rotation, warmup, sequences with automatic follow ups, double verification, the block list, the sales inbox, the lead extension and the analytics are all mechanical. Two features use models: an AI spintax button that varies subject, body and signature, and Autopilot, which turns a website into a profile, a lead list and a sequence. Neither is what the product is for.

The packaging confirms it, which is the reference test on this axis: none of the three tiers carries an AI line item or an AI credit, and every published meter counts emails per month, active contacts or lead credits.

One approval gate is published and it sits on the right feature. Autopilot, the most autonomous thing here, is described as turning a website into a profile, a lead list and a sequence on one approval, so the human confirmation is written into the feature description rather than buried in an enterprise tier. That is the right placement and it is a single clause.

Nothing states whether the approval can be switched off, what the approver actually sees, or whether any record is kept of what was generated against what was sent. Away from Autopilot the platform runs unattended by design, sending as many as 500,000 emails a month across an unlimited number of rotating accounts, and the pacing controls are throttles rather than oversight.

Two model backed features ship and the disclosure surface for both is empty. No provider, model family, version, inference location, retention position or accuracy figure is published for the spintax generator or for Autopilot. That second one matters more than its billing suggests: Autopilot reads a website and derives an ideal customer profile from it, which is an inference that decides who gets contacted, and no confidence measure or failure behaviour accompanies it. The one place a model vendor would have to appear is the published sub processor list, which was not read this pass and is the first thing to check.

Real named people saying real things, with almost no measurement behind any of it. The testimonials carry full names, roles and employers, including a co founder at Biizy reporting four times more positive responses and twice the monthly clients, and an operations manager at South East Scanning describing a specific contract closed with a named counterparty. A self hosted reviews page claims 36 reviews at 4.9, and the site claims 2,000 plus founders, agencies and sales teams.

Against that: no logos, no case study, no methodology anywhere, several quotes that read as support messages rather than outcomes, and reviews held on the vendor's own page rather than an independent platform. One point of genuine credit that runs the other way from the substitution pattern this index keeps finding: the sample dashboard carrying reply and bounce rates is labelled illustrative figures, so the vendor marks its own mock numbers as mock.

Shipped controls sit at the top of this band and the framing around them is the problem. Genuinely creditable: a block list that suppresses entire domains or individual addresses is a standard feature rather than an upsell, and a public opt out request page exists for people who want out of the data.

The vendor also names its regimes, claiming compliance with GDPR and with the California statute, and stating that European and British data moves under standard contractual clauses with a processing agreement available on request. What holds it here is that those are claims about the platform being used to send, not a position on the sending itself.

No consent requirement is placed on the buyer, no unsubscribe insertion, no regional filtering and no suppression obligation is described, and the regimes that actually govern unsolicited commercial mail go unnamed. The formulation to keep is the vendor's own: buyers are told to send cold emails without worrying because the platform holds a compliance badge. A processor's compliance does not discharge a sender's obligation, and this page reads as though it does.

The fullest data protection surface published by any vendor at this price point in the index, and it is not close. Eight documents are linked in the footer: terms, a privacy policy, a separate California privacy policy, a cookie policy, a data processing page, a security policy, a sub processor list and an accessibility statement, plus a standalone opt out request page.

A published sub processor list at a 23 dollar a month vendor is rarer than it should be; several enterprise platforms graded here do not have one. The transfer posture is stated rather than implied, with standard contractual clauses named for European and British data and a processing agreement available on request.

One engineering control from the security policy belongs here too and is specific in a way most privacy claims are not: pseudonymisation is applied deliberately to email verification data so that a breach cannot exploit it. Off the top band because the documents themselves went unread this pass, no data protection officer or Article 27 representative was located, and the compliance statements on the marketing pages are asserted rather than evidenced.

Two sources, both described plainly, and the plainness is most of the credit. The lead extension is documented honestly in the vendor's own FAQ: it requires the buyer's own Sales Navigator subscription, the buyer runs the search, and matching leads are exported with verified email addresses as a file. So the targeting decision and the licence both sit with the buyer rather than in a pool the vendor assembled.

Alongside it a database of 200 million verified leads is claimed, and there the provenance stops: no source, no lawful basis, no collection method and no named verification providers, on a product that says it uses two of them. What separates this vendor from most of the cohort is a published opt out request page, which is an actual removal route for the person in the data, and Clevenio, Crono, Cadivra and DM Faster were all marked down for having none. Off the band above because a request form is reactive where Adapt.io notifies the individual on first entry and publishes a self service lookup.

One real exposure, disclosed rather than concealed. The lead extension pulls leads with verified emails out of Sales Navigator search results, which is extraction from a platform whose terms do not permit it, and the vendor describes the mechanic openly in its FAQ rather than obscuring it. Three things keep it at this band rather than lower, and they are the Bindago distinctions. The buyer must hold their own Sales Navigator subscription, so no licence circumvention is being sold.

The extension runs in the buyer's browser under the buyer's own session, so the vendor never takes custody of the credential. And there is no account rotation, no proxying, no multi account stacking and no undetectability or evasion language anywhere on the site. The sending half carries no platform exposure at all, running through the buyer's own mailboxes.

Off the band above because no conformance position on the extraction is stated anywhere, and the account that gets restricted is the buyer's. Bindago sits above this within the band for putting the risk in writing in its own terms.

No training, tenancy or retention position accompanies either model backed feature. The spintax generator receives the customer's own email copy and Autopilot ingests a third party website, and nothing published says which provider performs the inference, where it happens, whether inputs are retained or whether one customer's material can inform anything served to another.

Partial credit is due from the security policy, which describes stewardship controls even though none of them is about models: refresh and access tokens encrypted at rest, passwords hashed with a named function, pseudonymisation of verification data, and a hosting arrangement under which the cloud provider holds no unencrypted customer data. A sub processor list is published and unread, and it is the document that would answer the model question.

Clean of impersonation and quiet about everything downstream of it. Nothing here invents a person, clones a voice, manufactures a location or performs social actions under the operator's name, and messages arrive from the buyer's own mailboxes at their own addresses. Three things hold it at this band. Spintax exists to vary otherwise identical copy so that a filter treats many messages as many messages, which is a signal aimed at the mail provider rather than at the reader.

Inbox rotation distributes one campaign across an unlimited number of sending identities, so nothing the recipient can see reveals the volume behind the note in front of them. And a person sitting in a 200 million record database learns none of this until the first message lands, with a request form as the only route out. Article 50 of the European AI Act goes unmentioned on a site that publishes marketing pages in five European languages.

Strong for the size, and the agent surface is what earns it. An MCP server ships as a listed feature rather than an enterprise upsell, marketed by naming the coding agents it is meant to drive, and it is the eleventh in this index. Beside it sits a documented cold email API with public documentation on its own subdomain, plus webhooks and integrations into an automation platform, a chat platform and two major CRMs.

The rest of the surface is unusually complete for a vendor at this price: sub workspaces, unlimited team seats, a one click migration wizard that imports campaigns, sequences, leads and schedules from a named competitor, a private community and a learning library. Off the top band because the CRM connections are webhook mediated rather than native two way sync, third party reviewers note the absence of deep native CRM integration, and there is no marketplace or partner directory.

Provider and country are both named plainly, which is more than most of this index manages. The security policy states that the services are hosted on Google Cloud Platform in the United States, that customer data stored there is encrypted at all times and that the cloud provider holds no unencrypted customer data.

Around that sits a coherent transfer posture rather than silence: a published sub processor list, standard contractual clauses named as the mechanism for European and British data, and a processing agreement available on request. For a United States hosted vendor selling into Europe that is a complete and honest answer, where the usual pattern in this corpus is to state neither the location nor the mechanism.

Off the top band because one region is all there is, no customer choice of location exists, no European hosting option is offered to buyers who need one, and the sub processor list was not read this pass.

A fuller enumerated control set than several far better funded vendors in this index, with no independent verification of any of it, which is the Distrobird shape repeating at a lower price. Published in detail: mandatory security training, password manager use with a unique credential per service, two factor authentication required and text message codes explicitly disallowed, least privilege role based access with removal on termination, multi factor authentication for production, network access control lists between security zones reviewed on a schedule, databases closed to all but internal addresses, tokens and passwords encrypted with named algorithms, internal review before deploy, penetration testing by independent third parties, regular vulnerability scanning with proactive patching, an ongoing bug bounty programme, three tiers of encrypted backup, and full logging of requests, sensitive user actions and support agent actions.

One control is more specific than anything comparable here: an account cannot be created using a password known to have appeared in a public breach corpus. Against all of it, no certification of any kind is claimed, no auditor, report, trust centre or attestation exists, and the document is dated 1 January 2024, so the most detailed security statement on the site is more than two and a half years old.

Three tiers, all priced, at 23, 47 and 179 dollars a month, with annual billing at twenty percent off and the exact yearly total published beside the undiscounted figure for each tier: 220.80 against 276, 451.20 against 564, and 1,718.40 against 2,148. Every meter is quantified per tier: emails a month at 5,000, 125,000 and 500,000, active contacts at 1,000, 25,000 and 100,000, and lead credits at 1,500, 5,000 and 10,000.

Credit top ups carry a published rate with a volume break, 80 dollars per thousand or 75 per thousand above five thousand, with plan credits stated not to roll over and purchased credits stated never to expire. A seven day free trial, cancellation from account settings at any time, and no contract. What earns the top band is what the pricing removes rather than what it discloses.

Sending accounts are unlimited on every plan and team seats are unlimited on every plan, so the two variables that make a cold email bill unpredictable elsewhere, per inbox billing and per seat billing, are both taken off the table, and the vendor leads with that. A section headed the whole invoice states that the subscription plus any top ups the buyer chooses is the entire bill, so the published number is the number the buyer pays, which is the actual test on this axis.

The vendor also publishes price comparisons against nine named competitors, benchmarking itself openly against the tier it undercuts. Two further items come from a direct read of the pricing page. First, the caveat the vendor volunteers against its own headline: asked whether email accounts are really unlimited, it answers yes with one honest caveat, and publishes the self hosted mailbox fair use allowance of ten, forty and one hundred and twenty by tier, so the single limit behind the unlimited claim is disclosed rather than discovered.

Second, a defect that does not move the grade: several competitor comparison pages published within the last fortnight quote a superseded price list, describing a forty seven dollar tier capped at thirty sending accounts and a two hundred and ninety seven dollar unlimited tier, neither of which exists on the current pricing page those same articles link to. Read as stale programmatic content trailing a repackaging rather than as a claim about the live offer. Recorded as unread rather than absent: the Platinum Service page listed in the footer with no price attached.

The architectural route out is genuinely good and the contractual half is undescribed. A documented public API, an MCP server and webhooks mean anything the platform holds can be pulled programmatically, the lead extension produces files by design, and campaign analytics can be shared by link. Cancellation is self service from account settings with no contract and no lock in. What is missing is everything after the account closes.

No post termination retention period, deletion timeline or deletion artefact appears on any page read, and no export function for leads, replies or campaign history is named in the feature list, so the file based route runs through the lead tool rather than the record store. The data processing page and the terms of service were not read this pass and are the single most likely place for the retention and deletion terms to sit. Re verify there before treating this as settled.

The strongest deliverability posture graded in this index since Cadivra, and one disclosure in it is a first. Warmup is described as ramping new mailboxes gradually with no fake engagement network. Reciprocal pools in which subscribers' mailboxes send each other synthetic mail to manufacture a sending reputation are the category standard, and this vendor names that practice and refuses it.

That is a specific architectural commitment of exactly the kind this axis exists to surface, and no other vendor graded has made it. Around it sits a real control set: double verification through two separate providers before a send, automatic gradual increase of daily volume on new campaigns, a block list at domain and address level, and reply and bounce reporting per campaign.

The free public toolkit is deliverability infrastructure in its own right, offering a sender authentication record generator, a spam word checker, a deliverability test, a header analyser and a bulk verifier with no signup. Off the top band on the volume architecture, which points the other way: unlimited sending accounts and rotation up to 500,000 emails a month, with no per mailbox daily rate, no stopping rule and no complaint threshold published. And spintax is sold explicitly as a way to help prevent messages going to spam, which is filter evasion rather than sending discipline.

Six buyer types, each with its own page rather than a line in a list: freelancers, founders, sales teams, marketing agencies, recruitment agencies and lead generation agencies. The homepage names the same set, and fifteen competitor comparison and alternatives pages place the product precisely against the tools a buyer would weigh it against. Marketing pages exist in five European languages. Off the top band because breadth is doing the work that definition should.

No headcount band, vertical, geography or ceiling is stated, the six audiences plus five languages plus a positioning built on undercutting a named rival describe a price led horizontal claim rather than a bounded market, and the only sizing figure published is 2,000 plus customers with no distribution behind it.