Type: Dark web search engine — .onion only
Access: Tor Browser required
Account required: No
Clearnet version: None
Indexes .onion sites: Yes
Content filtering: Unverified — partial claimed
No-logging claim: Yes — unaudited
Last verified: March 2026
What Is Excavator?
Excavator is a dark web search engine that operates exclusively as a .onion hidden service — there is no clearnet version and no alternative access method. It indexes .onion sites and returns search results through a clean, minimalist interface, claiming a strict no-logging policy for all search queries.
Its name suggests its purpose: digging through the dark web to surface content that other tools miss. In practice, Excavator occupies a middle position in the dark web search engine landscape — a larger index than Candle, a simpler interface than Haystak, less filtering than Ahmia and less coverage than Torch. It is most useful as a supplementary engine when primary search tools come up empty.
Onion Address
Note: Excavator has no clearnet version. This is the only access point. If the address does not load, verify the current address through Ahmia — Excavator’s address may rotate periodically.
How to Use Excavator
- Open Tor Browser with security level set to Safest
- Paste the .onion address into the address bar
- Enter your search query in the search box
- Browse results — each shows the .onion address and a brief description where available
- Verify any address from results against a second source before interacting
Honest Assessment of Excavator’s Claims
Excavator makes several claims about its privacy and security that deserve honest evaluation before trusting the service for sensitive searches.
| Claim | Verification Status | Practical Implication |
|---|---|---|
| No-logging policy | ❌ Unaudited — self-reported | Cannot be verified — treat as unconfirmed |
| Open-source code | ❌ No public repository linked | “Open-source” claim provides no practical transparency benefit |
| Encrypted search | ✅ Yes — Tor routing by default | This is a property of Tor, not specific to Excavator |
| No personalized ads | ✅ Verifiable — no ads visible | Accurate as observed |
| Eco-friendly servers | ❌ No documentation | Unverifiable marketing claim |
The practical conclusion: Excavator’s Tor-only architecture provides the same network-layer privacy as any .onion service — your IP is hidden from Excavator’s servers by default. Its specific privacy claims beyond this baseline are unverified. Use it as a search tool, not as a trust anchor for sensitive research decisions.
Excavator vs. Other Dark Web Search Engines
| Engine | Index Size | Filtering | Audited? | Best For |
|---|---|---|---|---|
| Excavator | Unknown | ⚠️ Partial — unverified | ❌ No | Secondary lookups |
| Ahmia | Moderate | ✅ Strong | ✅ Partially | Safe first search |
| Torch | Very large | ❌ None | ❌ No | Maximum coverage |
| Candle | Small | ❌ None | ❌ No | Quick lightweight search |
When to Use Excavator
Use Excavator when:
- Ahmia, Not Evil and Torch have all returned nothing useful for a specific query
- You want a simple .onion-only interface without the overhead of heavier search engines
- You need a backup engine when primary engines are experiencing downtime
Do not use Excavator when:
- You need verified, audited privacy guarantees for sensitive research
- You need comprehensive coverage — its index size is unknown and likely smaller than Torch
- You need content filtering — its filtering claims are unverified
The Recommended Search Engine Sequence
Excavator fits at the end of a logical search escalation sequence — try more established engines first and use Excavator when others have failed:
- Ahmia — filtered, reliable, Tor Project endorsed. Start here.
- Not Evil — broader coverage, partial filtering. Try if Ahmia has nothing.
- DuckDuckGo .onion — for clearnet context around a dark web topic.
- VormWeb — specifically for verifying whether an address is a phishing clone.
- Torch — maximum unfiltered coverage. Use with Safest mode.
- Haystak — deep coverage including paid tier for systematic research.
- Excavator — last resort when all above have failed.
Frequently Asked Questions
Is Excavator’s no-logging claim trustworthy?
It cannot be independently verified. Excavator has not published a technical audit, a transparency report or a warrant canary. The no-logging claim is self-reported with no external validation. Your IP address is hidden from Excavator’s servers by Tor’s routing — this is the verified privacy protection. Whether query content is logged is not verifiable from outside the service.
Does Excavator index .onion sites or clearnet sites?
Excavator indexes .onion hidden services — it is a dark web search engine, not a clearnet engine with a .onion address. Unlike DuckDuckGo’s .onion which returns clearnet results, Excavator’s results are .onion addresses. For clearnet search inside Tor, use DuckDuckGo’s .onion instead.
Why is Excavator’s index size listed as unknown?
Excavator does not publish statistics about its index size, crawl frequency or coverage. Unlike Torch which claims specific page counts, Excavator provides no documentation about the scale of its indexing operation. Index size can be roughly assessed by running specific queries and comparing result counts against known engines — but without official documentation, any estimate is approximate.
Is Excavator related to any other dark web service?
Excavator is an independent search engine with no publicly documented affiliation to other dark web services. It appears in the darknetking.com and h25.io directories as a listed search engine alongside Ahmia, Torch and others. Beyond directory listings, its operator identity and affiliations are not publicly known.
What should I do if Excavator returns results that seem wrong or outdated?
Dark web search engine indexes go stale quickly — sites go offline, addresses change and new sites launch faster than crawlers can track. If Excavator returns an address that does not load, the site is likely offline or has moved. Try Ahmia or Torch for the same query to compare results, and verify any found address through community sources before interacting with it.