Skip to content
HostRead

For independent podcasters and small networks

Find the sponsors your listeners already love. Get the pitch written.

An agent reads your listeners' taste in film, TV, music, food and books with Qloo, then ranks the brands they love.

Brand-safety exclusions Exclude

Or open an example deck

Are you a brand? Find podcasts whose listeners love you

One host's story, illustrative

Your topic isn't your audience.

  1. The usual way

    Dana hosts a small true-crime podcast with no ad-sales team. She asks a chatbot who should sponsor her and gets home security, mattresses and meal kits: the list every true-crime show gets. Her pitches go unanswered.

  2. With HostRead

    She types her show's name. The agent maps what her listeners love across film, TV, music, food and books, and finds a brand whose fans love the same things. The chatbot never mentioned it.

  3. What she sends

    A pitch email naming three tastes her listeners share with that brand's fans, a host-read script built from them, and a one-page media kit.

Every deck puts the plain-LLM guess next to the taste-grounded picks and marks the Surprise fits only the taste graph could find.

Plain LLM

Sees Dana's topic. Suggests the sponsors it gives every true-crime show.

HostRead with Qloo

Sees what Dana's listeners love, and which brands' fans love the same things.

  • Film
  • TV
  • Music
  • Food & drink
  • Books
  • Places
  • Podcasts

How the agent works

An LLM plans the run and calls tools. Every tool is backed by Qloo's taste graph or by deterministic code, and you watch each step as it happens.

  1. Resolve and profile the show

    Finds the show in the taste graph and maps what its listeners love across film, TV, music, food & drink, books, places and podcasts.

  2. Find candidate brands

    Ranks brands by affinity among your listeners, minus any categories you exclude for brand safety.

  3. Check the taste overlap

    Profiles each chosen brand's fans and measures how much their taste overlaps with your listeners'.

  4. Score in code, write grounded

    Code computes the fit. The pitch and script may only cite entities the API returned, and code checks every one.

How we score

Fit = 100 × (0.50·A + 0.35·O + 0.15·S)

A, audience affinity
Qloo's affinity of the brand among this show's listeners (/v2/insights, brand type, show as signal), min-max normalised across the candidate brands.
O, taste overlap
For the listeners' top 20 favourites in film, TV, music, books and food & drink, Qloo scores how much the brand's fans like each one. A weighted Jaccard compares the two sets of affinities, averaged across domains and min-max normalised within the deck.
S, surprise
1 when a plain LLM, given only the show's name and topic, did not suggest the brand. These brands get the Surprise fit badge.

Scores come from deterministic code, never from the LLM. A fit score is a relative ranking within one show's deck, not a prediction of revenue, CPM or conversion.

HostRead: sponsors your listeners already love