Two workflows. One creative director.
How Daryl works.
Daryl handles two things that used to eat your team's time: organizing your brand asset library intelligently, and finding the right asset for any given post — instantly.
Workflow 1
Your brand library, finally organized.
Connect your assets and Daryl does the rest. Every image is analyzed for visual context, brand alignment, mood, and message fit — automatically. No manual tagging. No hunting.
01
🏢Connect your brand
Tell Daryl who you are. Brand name, visual identity, target tone, and campaign context. He uses this to filter every recommendation through your brand lens — not generic AI taste.
02
📁Import your assets
Upload directly or connect Google Drive. Daryl accepts images in any format. Drop in a folder and he handles the rest — no file structure requirements, no renaming conventions.
03
🔍AI analyzes everything
Every asset is automatically analyzed by Daryl's vision model: dominant colors, visual mood, subjects, composition, brand alignment score, and semantic description — all indexed for instant search.
Then search in plain English.
Type anything: "warm lifestyle shot for fall campaign", "minimalist product with negative space", "energetic outdoor brand moment". Daryl's semantic search finds it — even if the filename is IMG_4821_final.jpg.
Search query
warm outdoor lifestyle, fall energy
→ 23 results ranked by brand alignment
Workflow 2
The right asset for every post.
Paste a draft post. Daryl reads the intent, tone, and audience — then returns ranked asset recommendations from your library, with a clear reason why each one fits this post.
01
Paste your post
Drop in your draft caption, blog excerpt, or ad copy. Daryl reads it as a whole — not just the keywords.
→
02
Daryl reads intent
He identifies the emotional tone, the message arc, the audience cues, and the visual energy the post is asking for.
→
03
Deep search
Daryl searches your entire asset library semantically — matching mood, brand fit, and visual context against your specific post.
→
04
Ranked + reasoned
You get your top picks, ranked by fit score, each with a plain-English explanation of exactly why it matches.
Example output
Your post
"Fall is here and we're leaning into it. Warm drinks, layered looks, and that golden-hour energy that makes every shot feel like a story. Which look is speaking to you this season? 🍂"
Daryl's top pick — 94% match
🍂
golden-hour-lifestyle-oct.jpg
Warm amber tones match the "golden-hour energy" directly. Subject layering aligns with "layered looks" copy. Seasonal brand score: 96/100.
Under the hood.
Daryl is built on a semantic search stack purpose-built for brand creative work.
Vision analysis
Gemini 2.5 Flash
Every asset gets a deep visual analysis pass: mood, composition, brand alignment, subject description, and semantic embedding.
Vector search
Pinecone
3072-dimensional embeddings indexed for semantic similarity — so "warm autumn energy" finds the right image even with zero keyword overlap.
Brand profile
Custom per team
Daryl builds a profile for each brand from your assets and context — meaning recommendations filter through your identity, not generic AI defaults.
Ready to put Daryl to work?
Connect your library. He'll have opinions ready in minutes.