TasteCast

Brand brief in, ranked AI characters out. Every claim cites its Qloo evidence, and close calls are called close.

Built for the Qloo Agentic Hackathon (2026).

• Live demo: https://tastecast.swanky-models.workers.dev

• MCP endpoint: https://tastecast.swanky-models.workers.dev/mcp (streamable HTTP)

• License: code MIT (LICENSE); character images CC BY-NC 4.0 (LICENSE-images); fonts SIL OFL 1.1 (public/assets/fonts)

TasteCast is for a studio that runs a small roster of original, disclosed AI characters (virtual models). A brand sends a brief; TasteCast asks Qloo which character the brand's audience actually shares taste with, shows its working, drafts the endorsement package, quotes a license, and then stops for people to approve. Every endorsement carries #ad · AI-generated character. The three characters (Itzel Robles, Mexico City; Ari Tseng, Seoul; Ingrid Brekkhus, London) are fictional and are not real users of any product.

Try it (judges, about 2 minutes)

1. Open the home page and pick one of the four demo briefs. Each card shows its stored verdict.

2. The brief fills itself, Qloo's /search candidates appear, and the exact match is confirmed (the human checkpoint). For North Kiln, Qloo lists two entries named "Le Creuset": the more popular one is preselected and the other is labelled.

3. Read the result: the verdict banner, the ranked characters (fit score /100, Qloo taste index in σ, audience similarity as a signed cosine), the Agent run timeline, the package, the license JSON (charged:false) and the evidence list. Every evN chip opens the Qloo call behind it.

Direct links: /brief?demo=0&confirm=1 … demo=3. Demo briefs are served from stored Qloo results and cost no live calls. Custom briefs make live Qloo calls under a small public budget (40 calls in total; per network 5 searches and 1 ranking a day). When it runs out, the site says so and points back to the demos.

Instant reference brands. Eleven well-known brands were run through Qloo ahead of time and stored, so a custom brief that uses one of them as the reference brand returns at once and makes no new Qloo calls. The brief page lists them, and the result carries a "Precomputed" note. Typed variants work (nike, Tajin, Stüssy, Blue Bottle). Leave the campaign city empty: a city adds an in-city check that was not precomputed. Verdicts below are for a generic brief and were not tuned.

Precomputing them cost 27 Qloo calls (Nike 6, Starbucks 5, Glossier, Aesop, Stüssy and Fender 4 each; the other five reused cached results), counted by a dry run before spending (scripts/instant-cost.ts).

Related MCP server: Sanbi ADK Agent MCP Server

Use it from an agent (MCP)

The same loop is a remote MCP server. Add it to any MCP client:

{ "mcpServers": { "tastecast": { "type": "http", "url": "https://tastecast.swanky-models.workers.dev/mcp" } } }

Tools: list_demo_briefs → resolve_brand → (your user confirms a candidate_id) → rank_roster → draft_package → quote_license. rank_roster refuses an entity that did not come from resolve_brand, returns the verdict, the ranking, the agent-run trace with evidence ids, and a view_url that opens the same run in the browser.

Smoke test with the official SDK client (uses a demo brief, no live Qloo calls):

npm install
node scripts/mcp-smoke.mjs https://tastecast.swanky-models.workers.dev/mcp 2
# rank_roster: Qloo decided (decided) lead 0.38 / needs 0.08; served stored; ...
#   agent run: plan → human → tool → tool → tool → tool → tool → check → check → check → check → tool → stop
# quote_license: US$11340 charged=false approvals={"character_operator":false,"brand":false}

Stateless and session-free (official @modelcontextprotocol/server SDK, JSON responses); works with 2025-era clients and the 2026-07-28 revision.

How Qloo powers it

Qloo is the decision-maker, not decoration. For each confirmed reference brand the agent calls:

1. /search to resolve the brand. A person confirms the entity before any insights call.

2. Audience (/v2/insights demographics): the brand's age and gender profile, compared with each character's seed profile as a cosine.

3. Taste index (/v2/insights, signal.interests.entities=<brand>, filter.results.entities=<all characters' seeds>): how strongly the brand's fans like each character's artists, brands, TV and film, against a baseline of six everyday brands, as z-scores. One query scores all three characters.

4. Context, no weight: shared taste tags and a house-brand fit check with explainability.

5. Package research: tags, artists and brands for the winning character's city and age bucket. Places that are shops or other brands are left out as shoot locations.

Fit score = 0.6 × taste index + 0.4 × (cosine + 1) / 2. "Qloo decided" needs a lead of at least 0.08 and the taste index must pick the same leader by z-score and by raw lift. Otherwise it is a close call, with the viable characters and their tradeoffs. Thresholds were set before the test runs and are not tuned per demo (see docs/milestone-2.md, docs/limitations.md).

Optional: Workers AI rewrites five copy lines once per package, in the background after the result is shown (the page swaps in the checked lines when they are ready). A checker rejects any rewrite that adds a name, a brand from the evidence, a number, or first-person/testimonial wording, or drops the disclosure; rejected lines keep the template text, and evidence ids are never written by the model.

Architecture

browser ──/api/*──┐                    ┌── Qloo Insights API (server-side key)
MCP client ─/mcp──┤  Cloudflare Worker ├── Workers KV (stored runs, cache, ledger, per-IP limits)
                  └──  Hono + service  └── Workers AI (optional checked copy rewrite)

• src/agent/ the loop: resolve, rank, decision, agent-run trace. src/content/ package, copy checker, license. src/qloo/client.ts cached client with ledger and budget guards. src/mcp.ts MCP tools over the same service as the web app. src/server.worker.ts Workers entry; src/server.node.ts local Node server.

• Details: docs/architecture.md, milestones 2–4, and docs/limitations.md.

Run it yourself

This repo contains no Qloo data: no response cache, no KV seed, no entity IDs. The live demo serves stored results from its own KV. To run locally you need your own Qloo key:

npm install
export QLOO_API_KEY=...     # server side only; never committed or sent to the browser
npm run precompute          # character profiles + everyday-brand baseline (about 40 Qloo calls)
npm run dev                 # http://localhost:8787 ; MCP at http://localhost:8787/mcp

Without a key the server still starts and Qloo-backed routes return a plain 503. Deploying to Workers: create a KV namespace, put its id in wrangler.toml, wrangler secret put QLOO_API_KEY, then npm run deploy.

Checks:

npm test             # 52 unit, route and MCP tests (no network)
npm run typecheck
npm run scan-public  # fails on cache/, .env, the key, raw Qloo JSON or UUID-shaped ids

Data and disclosure rules

• The Qloo key lives only in the server environment. Clients never see entity IDs; candidates use opaque tokens.

• Every live Qloo call is logged (timestamp, tag, path, status, latency) and counted against a total cap and a public sub-cap.

• License quotes are never charged and both approvals stay false until people sign off.

• The characters are AI-generated, labelled as such on every page and in every package.