Find better leads. Know who’s safe to reach.
Tuski combines AI-assisted company research with real delivery signals and a 0–100 quality score — so your team can find the right accounts, prioritize safer addresses, and protect sender reputation.
One number your team can act on
A 0–100 signal-based deliverability rating on every evaluated address. Tuski weighs provider and MX data, address patterns, disposable-domain and gateway signals, plus available domain-health context — so teams can sort, segment and review risk faster.
Classify what actually comes back
Tuski sends controlled probes, matches returned delivery-status notifications, and classifies observed outcomes with rules-based classification and pattern matching.
From list to evidence-backed results
Upload a CSV, follow progress as results arrive, and export actionable verdicts and scores for your next campaign or workflow.
How Tuski intelligence works
AI-assisted research finds and explains company fit. Transparent, signal-based analysis scores deliverability and classifies observed delivery outcomes.
AI-assisted company research
Search-capable AI gathers company context, firmographics and ICP evidence, with source URLs and confidence signals for review.
Signal-based deliverability scoring
Feature scoring combines MX and provider detection, address patterns, disposable-domain checks, gateway signals and available domain-health context into a sortable 0–100 score.
Real bounce evidence and classification
Probe sends are matched to returned delivery-status notifications and classified with rules and pattern recognition into invalid, full, blocked, routing, temporary and other outcomes.
AI-ready automation
Use API keys and MCP tools to validate lists, launch jobs, inspect status and export results from your stack or AI assistant.
Built for real revenue teams
“Before Tuski our bounce rate was wrecking deliverability. First clean run and we dropped under 1% — open rates jumped the same week.”
“We run every list through Tuski before a send. Same accuracy as the expensive tools, a fraction of the spend, and no-MX domains are free.”
Eight verdicts. Pay for the ones we reach.
Every address gets a stamp. You’re charged only when we send a controlled probe — domains we can’t reach are always on the house.
No bounce observed
The controlled probe completed without a returned bounce during the validation window.
Hard bounce
The recipient was rejected as invalid — a confirmed outcome pulled from your send list.
Domain can’t receive mail
No mail exchanger on record — nothing to send to, so it’s free.
Disposable address
Known temporary-mail domain detected before a probe is sent — free.
Fat-fingered domain
gmial, hotmial, outlok — no such mail server, so we never send. Free.
Possible catch-all
Multiple probes produced no observed bounces. Flagged as a heuristic signal for your review.
Mailbox at capacity
A returned status says the mailbox is full — a temporary failure that needs review.
Shared inbox
info@, sales@ and noreply@ patterns are marked so your team can segment them.
Built for your stack — and your AI agents.
Validate emails, check job status, retrieve or export results, and pause, resume or cancel work through API-key authenticated MCP tools.
{
"mcpServers": {
"tuski": {
"url": "https://tuski.ai/mcp",
"headers": {
"Authorization": "Bearer ${TUSKI_API_KEY}"
}
}
}
}{
"job_id": 1234,
"status": "QUEUED", // poll get_job_status
"count": 1 // verdicts arrive via get_results
}Run it from your AI assistant
Tuski speaks MCP. Wire it in with a single key — then queue checks.Watch statuses and collect validated results without ever switching windows.
Pay less. Validate more.
Subscribe for a daily allowance at the lowest rate — or switch to pay-as-you-go and keep your credits forever. You pay for every address we actually test — no-MX and unreachable domains are free.
Compared side by side
SIGNALS, EVIDENCE & AUTOMATIONWe stand behind every verdict.
Every evaluated address is checked against concrete signals, then controlled probe sends provide real delivery evidence when a mailbox returns a status notification:
A credit is spent only when we actually test an address with a controlled probe. Domains we can’t reach — no-MX, burners and other unreachable addresses — are never charged. Credits never expire. No small print.
Frequently asked questions
One credit is spent for every address we actually test — we send a real letter and read the mailbox's answer, so even a confirmed bounce is a result you paid to learn.Addresses we can't send to — no-MX / dead domains, burners and unreachable — are always free. Credits never expire.
Syntax, domain and MX checks are near-instant, but controlled probe sends run through rate-limited sending accounts — so observed delivery verdicts arrive over time rather than all at once.Results stream back as they land, and larger lists take longer to finish, up to hours for very large ones.
Tuski sends controlled probe messages and classifies observed delivery-status notifications. Results reflect real delivery evidence when it is returned, while the quality score remains a relative signal rather than a delivery guarantee.
Every evaluated row carries a signal-based quality score from 0 to 100. It weighs provider and MX data, address patterns, disposable-domain and gateway signals, plus available domain-health context so you can sort and review relative risk. It is not a delivery probability or guarantee.
Yes — connect over MCP with your API key and run validations from Claude Desktop, Claude Code or any MCP client: launch checks, watch statuses and pull scored results without leaving the chat.
Yes — that’s the point. Removing invalid, blocked and full mailboxes before you send keeps your bounce rate low, the single biggest factor in inbox placement.
Cleaning lists occasionally? PAYG — those credits never expire. Sending daily campaigns? A subscription gives a steady allowance at the cheapest rate — and you can keep PAYG credits as overflow.