SwipeHarbor MCP server

• Registry namecom.swipeharbor/swipeharborverified in the MCP registry

The registry name sits under swipeharbor.com. The registry only lets a publisher use that namespace after proving control of the domain through DNS or HTTP.

• Stateanswerschecked 2026-10-27
• Tools8
• Server calls itselfswipeharbor
• Server version1.0.0
• Protocol version2025-06-18

On 2026-10-27 the server answered initialize and tools/list without a sign-in. First answer in 74 ms.

• Whether its tools do the product's job has not been judged yet.
• Checked on 2 of the 7 days it takes to judge reliability.

Said by rules from the measurements above. The likely first call is a model's pick from the tool list; read its description before relying on it.

• INITE MCP scoreafter a weeka score needs 7 daily checks: 2 so far
• Answeredmeasuring2 of the 7 daily checks it takes to print a share
• Median first answer229 ms
• Space in an agent's context≈ 1,429 tokensthe definitions of all its tools (8), which an agent carries on every turn
• Tool definitionsB · 83 of 100
• Typed results0 of 8 toolsdeclared outputSchema: results come typed, read once

• Parameters without a description: in 8 of 8

A program checks the server once a day: initialize and tools/list, no model. Tokens are the definitions' length at about four characters a token. The grade averages six checks of every tool's name, description, schema and annotations; it says how well the tools are described, not whether they work. The score weighs what was measured and nothing else: answering (25%), doing the product's job (25%), definitions (20%), context (10%), use (10%) and how readily a client connects (10%). A part that was not measured is left out, and a hidden instruction caps the score at 40. The state of vendors' MCP servers: method and numbers →

• search_ads

  Search stored NewsBreak/Rumble ads. Results contain at most 12 ads. sort=newest (default) orders by last observation; sort=longest_running orders by recorded observation span in whole days. minDays=7|30|90 keeps ads observed over at least…

• get_ad_evidence

  Read one captured ad, its recorded category provenance, text evidence, and device context. Does not fetch advertisers or generate previews.

• get_funnel

  Read only the recorded destination path. Unresolved, failed and unvisited steps remain explicit; this tool does not navigate or recover hidden pages.

• compare_ads

  Return recorded evidence for two to four distinct ads for comparison. It does not infer competitor spend or performance.

• research_groups

  Research advertisers or destination domains on one platform (newsbreak, rumble or youtube). kind=advertiser groups by the recorded advertiser label; kind=destination groups by the declared or stored landing domain. Without group, returns…

• get_transcript

  Read the stored timed transcript of one Rumble video ad, 100 segments per page with start/end milliseconds. state shows when no transcript is ready; nothing is transcribed on request. Follow nextSegmentOffset for later segments.

• list_saved_ads

  List the caller's own saved ads, newest first, 12 per page, with the caller's collections. collectionId limits the page to one owned collection. Each item has the saved note and an ad summary, or ad_status=unavailable when the ad is no…

• save_ad

  Save one library ad to the caller's saved ads; optionally add it to one owned collection (collectionId from list_saved_ads) and set its note. Saving again is safe. A note replaces the existing note only when given. Use only when the user…