Starnus

Starnus positions itself as an “AI employee” for outbound sales. The current product is better understood as an orchestration layer than as a single chatbot or a single proprietary model: it analyzes a company and ideal customer profile, sources and scores leads, coordinates research and enrichment, creates outreach, sends email and LinkedIn messages, follows up, tracks replies and adjusts the next plan.¹²

The product’s main pitch is consolidation. Instead of asking a founder or small sales team to assemble a lead database, enrichment service, email sequencer, LinkedIn automation tool and reporting layer, Starnus tries to sit above those components and operate the workflow through natural-language goals.¹

Current first-party documentation lists 26 specialized agents. Named underlying services include Perplexity for web search, Smartlead for email automation, Unipile for LinkedIn automation, and Gemini 2.5 Flash for one image-processing agent, alongside integrations for HubSpot, Gmail, Slack, Notion, Google Workspace, Outlook, OneDrive, Airtable, Linear and X.² The public materials reviewed by AiToolMap do not identify one current foundation model that should be treated as the universal “Starnus model.” That is important: model benchmarks for Claude, Gemini or another provider cannot simply be transferred to the Starnus product.

This is a desk review. AiToolMap did not create a Starnus account, connect an inbox or LinkedIn account, buy credits, launch an outbound campaign, verify lead accuracy, test deliverability or measure meetings booked. Vendor performance numbers and user claims are reported as such, not converted into hands-on findings.

The current homepage describes a four-stage loop. First, Starnus analyzes a company website and extracts proposition, competitors, pain points and an ICP. Second, agents search and score leads. Third, the system prepares personalized outreach across email and LinkedIn. Fourth, it monitors replies and campaign metrics and adapts the following plan.¹

Starnus says six specialized agents can analyze a company site in parallel and that 26 agents handle lead sourcing, LinkedIn, email, follow-up and monitoring. The homepage also emphasizes a human approval step before outreach sends.¹ The managed-service page is even more explicit: nothing goes out before sign-off, and users review prospect lists and messages before campaigns launch.⁵

This human-in-the-loop design is a meaningful strength for an outbound product. An autonomous sales agent can create reputational and legal problems at machine speed if it targets the wrong people or invents claims. Starnus’s current public design at least recognizes that sending should be gated by approval.

The broader agent catalog shows that the architecture reaches beyond sales. Starnus exposes agents for PDFs, Excel, Gmail, Slack, Notion, Google Docs/Sheets/Drive/Calendar, Outlook, OneDrive, Linear, Airtable and X.² That breadth makes the product closer to a business automation layer with outbound specialization than a conventional single-purpose AI SDR.

Starnus repeatedly compares itself with stacks built from Apollo or Clay plus a sending tool. The argument is intuitive: the more systems a team must configure and hand data between, the more operational overhead it inherits.¹

The strongest independent-ish user evidence points in the same direction, although the samples are tiny. Product Hunt currently shows 4.7/5 from three reviews. Reviewers praise keeping prospecting and outreach together, reducing tool juggling and automating follow-up; one reviewer specifically says the lookalike/prospect-selection stage is interesting but wants more visibility and control over why accounts were selected and clearer onboarding.⁸

Trustpilot’s four-review sample is more polarized. Positive reviewers say Starnus simplified a multi-tool outbound setup, saved time and handled lead finding, messages and follow-ups. The single negative reviewer alleges a spam-originated contact, recurring $60 billing, poor support/onboarding and login/password problems.⁴ Trustpilot itself warns that the company has not recently invited customers and that the sample may not be representative. One positive review uses the same name as Starnus co-founder/CEO Khashayar Mansourizadeh; AiToolMap cannot independently establish the identity or relationship of that reviewer, so it treats the entire sample with additional caution.⁴

These signals are useful for themes, not for proving ROI. There is no credible panel evidence showing that Starnus consistently produces better-fit leads, higher reply rates or more booked meetings than a well-run conventional sales stack.

The current homepage advertises “3× better-fit lead selection,” “8× faster outbound execution,” a 30-second ICP analysis and 200+ leads per week.¹ Those numbers are not accompanied in the material reviewed here by an independent benchmark protocol, sample definition or reproducible comparison.

The managed-service page says most clients can have a first campaign live within three to five days and describes weekly optimization and reporting.⁵ Again, that is a service description, not an independent outcome study.

Starnus’s own Terms are more careful than its marketing. They explicitly state that neither self-serve nor managed services guarantee any number of leads, meetings, replies, conversions or revenue results; estimates and projections are aspirational.⁷

AiToolMap therefore does not use Starnus’s 3×, 8×, 200+ leads/week or meetings claims as rating inputs.

Starnus also sells a managed service in which a dedicated outbound manager operates the platform on the client’s behalf. The current page says the service covers ICP research, lead sourcing, personalized LinkedIn/email outreach, campaign management, reply handling, meetings and weekly reporting, starting from $900/month plus a Pro+ or Ultra plan.⁵

This is not merely “more credits.” It turns Starnus into a hybrid of software and an outbound agency. For a founder with no SDR function, that can be attractive because it transfers execution overhead to a human-plus-agent team.

It also changes what should be compared. A managed-service buyer should benchmark Starnus against agencies and outsourced SDR services, not only against software such as Apollo or Lemlist. The service’s Terms say a separate Managed Service Agreement can govern the engagement and explicitly place responsibility for campaign cooperation and legal compliance on the client.⁷

Automated outbound is legally and operationally sensitive. Starnus’s Terms make the user responsible for ensuring target audiences and communications comply with GDPR, the ePrivacy Directive, CAN-SPAM, CASL and other applicable rules. Users must also comply with third-party platform policies, and Starnus disclaims liability for restrictions or penalties imposed by those platforms.⁷

Starnus says it uses commercially reasonable deliverability/compliance practices such as opt-outs, domain warm-up and volume management, but explicitly says this is not legal advice.⁷

That allocation of responsibility is important. A user cannot treat “the AI selected these people” as a compliance basis. ICP selection, lawful basis, suppression/opt-out handling, sender identity and claims in outreach remain governance decisions for the business.

The current G2 AI Marketplace page lists Starnus at €50/month for Pro, €125/month for Pro+, €250/month for Ultra and managed service from €900/month.³ Starnus’s current managed-service page independently shows a $900/month starting price, with a Pro+ or Ultra platform plan added.⁵

A Starnus pricing page is still indexed with the same euro self-serve ladder and describes 60,000, 150,000 and 300,000 monthly credits across Pro, Pro+ and Ultra. However, during this review cycle the live `/pricing` URL returned 404 when opened directly. AiToolMap does not treat the indexed search text as a substantively read pricing page under v1.8.

The Terms provide the most reliable current billing mechanics. New accounts receive 12,000 trial credits, but a valid payment method and selection of a paid plan are required before using them. The trial lasts three days; if it ends naturally, the selected paid plan activates and the card is charged automatically. Credits reset each billing cycle, unused credits do not roll over, and fees/credits are generally non-refundable except where law requires otherwise.⁷

That current contractual setup differs materially from Starnus’s Product Hunt launch offer, which advertised a 14-day trial without a credit card and much lower launch pricing.⁸ The launch information is now stale and should not be used for purchase decisions.

The pricing proposition can still be attractive if Starnus genuinely replaces several paid tools. But value should be judged against actual credit consumption, connected-account limits and outcomes—not against the sum of every competing product’s full subscription price.

Starnus Technology B.V. identifies itself as the data controller for account and platform data, with an Eindhoven address and Dutch/GDPR framing.⁶

The Privacy Policy says Starnus can process account and billing information, project and agent configuration, API keys, OAuth tokens, connected-app content, prompts, uploaded files/images, outputs, usage logs and approximate location.⁶ When users invoke integrations or third-party agents, Starnus can transmit prompts, context, connection tokens and the minimum relevant data to those providers.⁶⁷

For managed service, the policy distinguishes roles: the client is controller for prospect data and Starnus acts as processor on the client’s instructions. Starnus may use lead-enrichment/verification providers, email infrastructure, LinkedIn automation tools, cloud providers and AI model providers as subprocessors. A GDPR Article 28 DPA is available on request where required.⁶

The policy says Starnus does not sell personal data, uses Standard Contractual Clauses or adequacy mechanisms for relevant international transfers, and deletes/anonymizes account personal data within 30 days after deletion except legally required records. Security logs may be retained up to five years.⁶

This is reasonably detailed disclosure for a young sales automation product. The privacy risk is not that Starnus hides the existence of integrations; it is that the product’s value depends precisely on giving an orchestration layer access to many business systems and prospect-data flows. Teams should connect only the scopes they are prepared to delegate.

Starnus’s Privacy Policy says it uses encryption in transit and at rest, role-based access controls, regular penetration testing and internal security policies.⁶

AiToolMap did not locate a fixed-panel independent security audit or a current public SOC 2/ISO certification for Starnus in the materials reviewed. The absence of a certification is not proof of weak security, but it matters for organizations that connect corporate inboxes, CRMs and prospect databases to an autonomous workflow.

The Terms also warn that third-party providers’ output may be inaccurate or unlawful and that connected platforms can impose their own restrictions or penalties.⁷ That is particularly relevant to LinkedIn automation, email deliverability and prospecting data sources.

The current product framing gives users a significant approval role. The homepage says every message is shown before it sends; the managed-service page says prospect lists and messaging require sign-off.¹⁵

At the same time, Starnus is designed to keep executing after a plan is approved. It can source data, enrich records, use external tools, send messages, monitor interactions and schedule future work. That makes observability and rollback more important than in a normal writing assistant.

The small Product Hunt sample raises exactly this issue: one reviewer wants more explanation of why particular prospects were selected before outreach begins.⁸ That is a useful product-design signal. In autonomous outbound, “why this lead?” is not merely an analytics feature; it is part of trust, brand safety and compliance review.

Starnus is most interesting for founder-led B2B companies, small GTM teams and organizations without dedicated sales operations that already know they want outbound but do not want to become experts in several separate tools.

It can also make sense for teams considering an outsourced SDR/agency model, because managed service provides a human operator on top of the agent system.

It is a weaker fit for organizations that require mature independent security attestations, a large verified-user evidence base, highly granular control of every data source and outreach stage, or a procurement process that cannot tolerate short auto-converting trials and rapidly changing product surfaces.

Before connecting production accounts, a team should verify the current paid tier, trial conversion date, credit economics, cancellation path, DPA/subprocessor requirements, LinkedIn/email policy exposure and approval controls. For regulated or high-reputation outreach, test with a constrained audience first.