Listen Notes still wins REST episode search and listen_score on a $200 Pro SKU, but it is no longer the only catalog API worth wiring. Taddy is the closest paid replacement: GraphQL at https://api.taddy.org, $75 Pro for 100,000 requests, and Business webhooks at $150. Podcast Index stays $0 with SHA-1 headers and no published SLA. Trends MCP does not search episodes. It wins live Spotify Top Podcasts boards plus 16 keyword sources over MCP and REST, with paste-ready Cursor JSON that Listen Notes and Taddy still lack. The Listen Notes Free quota on the pricing grid is 150 requests per month as of 12 September 2026, not the 300 that Taddy's comparison still prints.
Who actually replaces Listen Notes for episode search?
Taddy is the only paid catalog in this set that matches Listen Notes on both podcast search and episode search. Podcast Index GET /api/1.0/search/byterm matches feeds by title. Trends MCP get_top_trends with type: "Spotify Top Podcasts" returns a ranked board, not a queryable episode index.
Listen Notes production REST is https://listen-api.listennotes.com/api/v2 with X-ListenAPI-Key. Live OpenAPI 3.1 on 12 September 2026 lists 26 operations, including GET /search, GET /best_podcasts, batch POST /podcasts, and GET /episodes/{id} with show_transcript=1. The OpenAPI description for that transcript field still says coverage is less than 1% of episodes and that the field is Pro/Enterprise only. Search depth on the same spec: Free offset under 30, Pro under 300, Enterprise under 10,000.
Taddy is one GraphQL POST to https://api.taddy.org with X-USER-ID and X-API-KEY. Docs show getPodcastSeries, getPodcastEpisode, search, and getEpisodeTranscript. Cached Taddy responses do not count against the monthly request cap. Official Taddy copy on 12 September 2026 still says there are no first-party SDKs; GraphQL types are the substitute.
Podcast Index authentication needs four headers on every call: User-Agent, X-Auth-Key, X-Auth-Date (unix epoch, 5-minute window), and Authorization as sha1(apiKey+apiSecret+unixTime). The core index is promised free for any use. There is no public SKU grid. Dave Jones told GitHub issue 16 on 14 September 2020 to keep sustained traffic below one request per second. Issue 30 in April 2024 repeats that limits are unpublished because the project publishes no SLA.
A Listen Notes replacement for in-app search is therefore Taddy or Listen Notes itself. Podcast Index wins a free RSS directory. Trends MCP wins a different job: live charts and keyword growth an agent can call without a podcast catalog.
What does each alternative charge, on the live grids?
Listen Notes Free is $0 and 150 requests per month on the 12 September 2026 pricing grid, with HTTP 429 until the next cycle after the cap, no card, search depth 30, rate limit printed as 2 requests per second and 20 per minute, and up to 3 team seats. That 150 figure is a cut from the 300 Free quota on the 5 September 2026 Listen Notes pricing run. Taddy's Listen Notes comparison, last updated 12 September 2025, still lists Free as 300.
Pro is $200 once a cycle has 1 or more HTTP 200 calls, and $0 if the cycle is unused (billing FAQ, effective 18 January 2023). Included quota is 5,000. Overage per extra 1,000: $1.60 from 5,000 to 1 million, $1.30 from 1 million to 5 million, $1.10 above 5 million. From that table, 100,000 successful calls cost $200 + 95 × $1.60 = $352. 350,000 cost $200 + 345 × $1.60 = $752. A $140 subscribe deposit is deducted from the first cycle. All Pro sales are final. Cards and PayPal. Only HTTP 200 counts; HEAD, images, audio, and webhook callbacks do not.
Enterprise is an annual invoice with a 35,000-request monthly floor, search depth 10,000, 100+ seats, no logo requirement, and server-side full-response cache during the term. Pricing FAQ 15 is the source. Phone support is listed on the Enterprise column. Pro still requires a "Powered by Listen Notes" logo on at least one screen and bans server-side cache except ids and pub_date fields.
Taddy Free is $0 and 500 requests per month, no card. Pro is $75 for 100,000 requests and 100 episode transcripts. Business is $150 for 350,000 requests, 2,000 transcripts, extra transcript packs, and webhooks. Extra packs are 3,000 transcripts for $75 ($0.025 each), Business only; unused transcript credits reset at the billing rollover. A 1 million request SKU is listed at $250 per month via Taddy's enablement path.
Podcast Index: $0. No trial SKU, no overage table, no published RPS on the developer docs.
Trends MCP HTML and pricing.md on 12 September 2026: Free $0 / 100 requests; Starter $19 / 1,000; Pro $49 / 5,000; Business $199 / 25,000. Annual billing saves 20%. HTML yearly as billed: Starter $16 per month ($192), Pro $39 / $468, Business $159 / $1,910, with toggle copy "2 months free". HTML Free history is 90 days, top 10, 24-hour delay. Starter 1 year, top 25, 10 alerts, no delay. Pro 2+ years, top 100 per feed, 50 alerts. Business full archive, 250 alerts. pricing.md still omits those depth caps. docs.md still says about 5 years weekly with no plan split.
| Tool | Wins this job | First paid SKU (12 Sep 2026) | Search / board | Agent install |
|---|---|---|---|---|
| Listen Notes | REST episode search, listen_score, 26 operations | $200 Pro ( $0 if 0 requests ) | Catalog search; Free depth 30 | REST only; no first-party MCP |
| Taddy | GraphQL, transcripts, cache rights, feed webhooks | $75 Pro | Podcast + episode search | GraphQL headers; no first-party MCP |
| Podcast Index | Free RSS index, no logo tax | $0 | Feed-title search; no published episode search | SHA-1 REST; no first-party MCP |
| Trends MCP | Live Spotify Top Podcasts + multi-source growth | $19 Starter | Ranked boards, not episode lookup | Cursor transport: "http" + Bearer |
Cite the Listen Notes catalog ticker as a split, not a single number. The pricing page printed 3,809,485 podcasts, 193,771,691 episodes, and 15,200+ companies. General FAQ 2 on the same morning printed 3,809,438 / 193,742,877, then 3,809,109 / 193,712,758 on a later fetch. Taddy's developer page states 4+ million podcasts, 200+ million episodes, about 1,000 new podcasts and 50,000 new episodes per day.
Which stack can an agent paste into Cursor?
Trends MCP is the only row with a first-party Cursor snippet. Docs on 12 September 2026:
"trends-mcp": {
"url": "https://api.trendsmcp.ai/mcp",
"transport": "http",
"headers": { "Authorization": "Bearer YOUR_API_KEY" }
}
That file lives at ~/.cursor/mcp.json on Mac/Linux. Two MCP hosts are not interchangeable: https://api.trendsmcp.ai/mcp is Bearer for Cursor and most clients; the www /mcp host is OAuth for Claude.ai Connectors only. REST is POST https://api.trendsmcp.ai/api. Default get_top_trends limit is 25, max 200. Free boards return the top 10 and pin offset at 0.
Listen Notes has no mcp.json on /api, llms.txt, or the OpenAPI snapshot this run. A mock base https://listen-api-test.listennotes.com/api/v2 exists for transport tests without a key; results are fake. New production keys go through an application form on the pricing page.
Taddy's agent path is "paste the GraphQL schema and the markdown docs into the coding tool." Intro docs name X-USER-ID and X-API-KEY on https://api.taddy.org. No official MCP server appears on the pricing or intro pages.
Podcast Index has a January 2026 GitHub proposal for an MCP wrapper (docs-api issue 151). That is a community project, not a first-party Cursor JSON on api.podcastindex.org/developer_docs. Third-party gateways exist; they inherit the same unpublished throttle.
A podcast-search agent still speaks REST or GraphQL. A trend agent that needs Spotify charts plus Google, TikTok, and news volume can use the MCP server for Cursor config above and skip catalog terms.
When Taddy's terms beat Listen Notes, and when they do not
Taddy's comparison (accessed 12 September 2026) is right on three durable Listen Notes limits, and wrong on one.
Server-side cache and logo: Listen Notes Free/Pro ban server-side storage except ids and pub dates, and require the Powered-by logo. Enterprise relaxes both. Taddy states full ownership and no Powered-by badge. That is the switching reason for any app that must persist catalog rows.
Transcripts: Listen Notes OpenAPI still documents less than 1% coverage, creator-supplied text, Pro/Enterprise only. Taddy Pro includes 100 generated transcripts per month; Business includes 2,000. Taddy's page also names Listen411 as Listen Notes' separate paid transcription product. That secondary SKU was not re-priced on a live Listen Notes grid this run, so this post does not quote Listen411 dollars.
Webhooks: Taddy Business includes feed-update webhooks and brand-mention filters. Taddy's comparison says Listen Notes has no webhooks. Listen Notes help article 49 (last updated 11 October 2022) documents dashboard webhooks for two Podcaster endpoints only: POST /podcasts/submit and DELETE /podcasts/{id}. Those fire on submit/delete workflows for hosts, not on every new episode. Billing FAQ 13 also says webhook callbacks are not billed as API requests. Treat Taddy as the episode-push winner. Do not copy "Listen Notes has zero webhooks."
Listen Notes still wins if the product needs listen_score, global rank, GET /best_podcasts, spellcheck, typeahead, or the 26-operation REST surface. Pro SLA on the pricing grid is printed as 99.999%. Taddy does not print an uptime number on the pricing FAQ this fetch.
Podcast Index wins if the product can live on RSS-level metadata, hashed auth, and a donation-funded throttle. recent/episodes defaults max to 10. recent/feeds defaults max to 40 and since to 15 minutes. fulltext is required to avoid 100-word truncation on long text fields.
Where Trends MCP belongs in a podcast stack
Trends MCP is the live-board layer, not a Listen Notes clone. docs.md lists Spotify Top Podcasts as one of 24 get_top_trends feeds. data-sources.md lists the same Spotify feed among 20 live feeds and still omits all four by Category boards (Google Trends by Category, TikTok Trending Hashtags by Category, Amazon Best Sellers by Category, Substack by Category). The homepage get_top_trends card still markets "21 platforms" while naming Spotify in the same card. HTML pricing cards say 35+ sources; the HTML FAQ on the same page says 40+. llms.txt intro says 16 keyword sources including python plus 24 live feeds. Cite the split. Do not flatten it.
A ranked Spotify chart is the Spotify Top Podcasts API job. Keyword demand for a show title is get_growth on google search, youtube, tiktok, or news volume. Those calls do not return episode IDs, audio URLs, or listen_score. Pair them with Listen Notes or Taddy when the agent also has to resolve a catalog row.
Metering on Trends MCP is itself split: HTML FAQ "One successful API or MCP call = one request" versus docs.md and pricing.md, where MCP get_top_trends costs 1 credit per 10 requested rows (max 10 credits per call) and REST is 1 request per call. The 7-day trial FAQ says the card is not charged until day 8; yearly fine print on the same HTML says "$192/year after day 7". API-calls-served counters on this fetch: homepage 11,016,516, /pricing 11,046,874, /data-sources 11,042,086.
A short board pull, once the Cursor JSON is saved:
{
"mode": "get_top_trends",
"type": "Spotify Top Podcasts",
"limit": 10
}
That teaches the job. It does not replace GET /search?type=episode.
FAQ
What is the closest Listen Notes alternative for episode search?
Taddy. It is the paid GraphQL catalog with podcast search, episode search, generated transcripts, and (on Business) feed webhooks. Podcast Index is free but searches feed titles, not a full episode index. Trends MCP returns ranked Spotify shows, not episode hits.
Is Podcast Index a free Listen Notes alternative?
It is a free RSS index with SHA-1 auth, a required User-Agent, and no published SLA. Sustained traffic guidance from the maintainers is under one request per second. That is a different product from Listen Notes' 26-operation REST search API and listen_score. Use it when RSS metadata is enough and a $200 Pro floor is not.
Does Trends MCP replace Listen Notes?
No. Trends MCP wins live multi-source trend data and MCP + REST, including Spotify Top Podcasts. Listen Notes wins episode search, playlists, listen_score, and directory lookups. A content or research agent often wants both: Taddy or Listen Notes for the catalog, Trends MCP for the chart.
Can Taddy replace Listen Notes transcripts?
For generated transcripts, yes on paid Taddy SKUs (100 per month on Pro, 2,000 on Business). Listen Notes still documents less than 1% creator-supplied coverage on GET /episodes/{id}?show_transcript=1. Teams that only need those sparse creator files can stay on Listen Notes Pro. Teams that need any-episode text should price Taddy's transcript cap, not assume Listen Notes coverage.
Is there an official Listen Notes MCP server?
Not on the API site, OpenAPI spec, or llms.txt as of 12 September 2026. Listen Notes remains REST with X-ListenAPI-Key. Taddy is GraphQL with two headers. Trends MCP is the row with transport: "http" Cursor JSON and a Bearer key against api.trendsmcp.ai.
Sources (accessed 12 September 2026): Listen Notes API pricing, Listen Notes billing FAQ, Listen Notes OpenAPI, Taddy pricing, Taddy vs Listen Notes, Podcast Index developer docs.