Replit Agent integration · MCP · security-scanner reviewed
Use Etymolt in Replit Agent.
Narrow surface. Scanner-reviewed. Verdict in two seconds.
Add Etymolt as an MCP server from the Integrations pane. Replit's security scanner reviews the tool definition; once approved, your Agent can verify any name candidate against live USPTO, RDAP, and 14 social-handle probes.
Sub-3s p95 · 12.7M US marks · 14 social platforms · free for five
What's available today
Why this, why now
The scanner is the feature. Not the obstacle.
Replit's security review catches MCP servers that exfiltrate data or take destructive actions. Etymolt's tool surface is structurally narrow — one read-only name per call, no filesystem, no shell, no persistent session. The narrowness is by design; it makes the verdict primitive a reliable building block for any Agent running in Replit, not a privilege-creep risk.
Three steps
Integrations → MCP Servers → Add.
I · Open Integrations
Side pane.
In your Repl, open the Integrations pane and choose MCP Servers.
II · Add Etymolt
Paste the URL.
Click Add Server. Paste this URL.
https://api.etymolt.com/mcpNo API key required for the free tier.
III · Scanner clears
Wait briefly.
Replit's security scanner reviews the tool definition. Once cleared, etymolt appears in your Agent's tool list. The narrow surface (one name per call, read-only) means review is fast.
Alternative · JSON-config equivalent
For Repls that surface raw MCP server config, paste this block. Functionally identical to the URL path.
{
"mcpServers": {
"etymolt": {
"command": "npx",
"args": ["-y", "@etymolt/mcp-server"],
"env": { "ETYMOLT_API_KEY": "(optional — 5 free per IP)" }
}
}
}How it fires
Replit Agent reaches for Etymolt on every naming intent.
I
You prompt the Agent.
"Scaffold X" or "rename this Repl" — any prompt that signals naming intent. Replit Agent proposes a candidate.
II
Agent calls verify_brand_name via MCP.
Etymolt fans out across USPTO + RDAP + 14 social probes in parallel. Sub-3s p95. The Agent surfaces the verdict in its reply with the disclaimer quoted verbatim.
III
Verdict drives the next step.
PROCEED candidates can be committed. ABANDON candidates do not get baked in. ITERATE and DUE_DILIGENCE candidates surface with the blocker called out so the Agent can iterate before deploying.
You Replit Agent Etymolt (MCP) │ │ │ │ "scaffold X" │ │ │ ───────────────────▶ │ │ │ │ scanner pre-cleared │ │ │ ─────────────────────▶│ │ │ call: verify_brand_name │ │ │ │ │ ◀─── verdict + score ─┤ │ ◀── verdict + ── │ │ │ disclaimer │ │ │ (verbatim) │ │
The validation stack
Five axes. One signed verdict.
Trademark
USPTO · TTAB · UKIPO · Madrid
Domain & handles
RDAP · WHOIS · 14 platforms
Cultural
7 languages · 20 markets
Linguistic
1.4M-name distinctiveness corpus
Sound symbolism
12 perceptual axes · TTS round-trip
What we are NOT
We don't generate names. We validate them.
Replit Agent proposes candidates. Etymolt verifies them against live ground truth across five axes. The two products sit cleanly on different sides of the workflow.
FAQ
Common questions.
How do I install Etymolt in Replit Agent?
Open your Repl, choose Integrations from the side pane, select MCP Servers, click Add Server, and paste https://api.etymolt.com/mcp. Save. The etymolt server appears in your Agent's tool list once Replit's security scanner has reviewed the tool definition.
What does Replit's security scanner do, and is it a problem?
Replit reviews every MCP server's tool definitions before letting an Agent invoke them — this catches servers that exfiltrate data, take destructive filesystem actions, or expose unsafe shell access. Etymolt's tool surface is narrow (a single read-only name string per call), so the scanner clears it cleanly. Once approved, the Replit-vetted signal benefits Etymolt's broader trust posture.
Why use a verification layer inside a Replit prototyping environment?
Replit Agent often scaffolds projects, repos, and deploys for users who are exploring an idea quickly. The naming moment shows up early — “what should I call this Repl?” — and a name that looks good in chat can carry a live USPTO collision or a taken handle on the platform the user plans to publish to. Catching it here is far cheaper than catching it after the user has bought a domain or signed up for ads.
Does this generate names? Does it compete with Replit Agent?
No. Etymolt does not generate names. We validate the names Replit Agent's model already proposes. Replit Agent makes the candidates; we return the verdict. The two products sit cleanly on different sides of the workflow.
Does Etymolt see my Repl content?
No. When Replit Agent invokes verify_brand_name, only the name string and an optional category hint travel to api.etymolt.com. No Repl code, no Agent chat history, no Replit account identifiers. We log: name, timestamp, hashed source IP. Retained 90 days.
Is Etymolt listed in Replit's MCP directory yet?
Submission is queued (board memo: item #29, deferred-30d). Until listed, the self-serve install path above works identically. Listing reduces install friction for non-developer users; the integration mechanics are the same.
What does this cost?
Five verdicts per IP are free. Signed-up users get ten more. After that, ten dollars buys ten more. No subscription. Your Replit subscription is billed by Replit — separate line item.
Bureau Model · clearance signal · not legal advice
Bring verdicts into Replit.
MCP · one URL, every Repl · see all surfaces →
Clearance signal, not legal advice. Confirm with trademark counsel before adopting a name in commerce. Data sources have stated freshness windows; refer to coverage_caveat per jurisdiction.
NPM · @etymolt/mcp-server