Competitor analysis · workshop readout

the market the wedge

Stellar × Nightborn · Workshop
The competitive landscape, and what it means for the “Influencer Suggestions” bet.

AI Influencer Discovery.

Executive summary

What the
market is
telling us

The market already moved

01

Natural-language search, brand-fit scoring, lookalikes and campaign learning are table stakes across every serious platform. Not a differentiator.

The premise checks out

02

The three-phase mechanism — Search → Profile → Learning — mirrors what the leaders ship. Phase 1 is the architecture Modash sells as an API.

Plumbing is commodity

03

No major vendor guarantees translation accuracy. A measured eval layer is the genuinely defensible piece — and the one Stellar doesn’t have today.

The real moat is Phase 3

04

Phase 1 is near-parity with what Stellar already ships. Performance-weighted learning is what rivals treat as their moat too.

The core tension

The plumbing is now
commodity.
Accuracy is not.

Every leader ships the same search, scoring and lookalike features — the engine and the data provider are things you rent. No one publicly guarantees that a brief becomes the right shortlist. A measured guarantee on real briefs is the defensible wedge, and the piece Stellar has to build to sell the story.

Market context

The shift is done, not coming

0%

rank AI-driven creator matching their single biggest 2026 priority (IMH Benchmark).

0%

of brands already use AI for creator discovery — the most mature AI use case in the stack.

0%

of marketers use AI to scale discovery, workflows and analytics (Aspire 2026).

Text-only filters are now the visible gap. This is a catch-up move — not a moonshot.

The checklist a prospect uses

Four capabilities now define the category

Every leading platform converges on the same feature set. Nightborn’s Phase 1/2/3 maps onto capabilities 1, 2 and 4; lookalike is cheap to bolt on.

01

Brief → shortlist

Describe what you want in plain text (or an image/URL), get a ranked list.

02

Brand-fit scoring

A score rating how well a creator matches the brand and how authentic the audience is.

03

Lookalike discovery

“Find 20 more like this one” — expand a proven creator into new tiers and markets.

04

Learning from campaigns

Re-ranking and suggestions weighted by realised performance. The stickiest layer.

How they actually built it

Two architectures, very different costs

A

LLM → filters
translation

  • A low-cost LLM reads the brief, outputs structured JSON mapped to existing filters.
  • The platform’s normal search runs, pre-filled. You reuse the engine you already pay for.
  • Used by Nightborn’s Phase 1, Stellar’s current “AI Search”, Upfluence, IMAI. Cheap: the AI is a translator.
B

Embeddings /
vector search

  • Creator content (images, video) becomes vectors; queries match by similarity.
  • Finds aesthetic matches (“cottagecore”, “boho kitchen”) that keyword filters miss.
  • Used by StoryClash (OpenAI over billions of images), Modash, HypeAuditor. Heavier: real “vibe” matching.

Nightborn quotes Architecture A for Phase 1 — correctly the cheaper path. The differentiator sits on top: measured accuracy, not the plumbing.

The field · six ways in

Who Stellar is measured against

The head-to-head

StoryClash / Kolsquare

European powerhouse (Kolsquare acquired StoryClash, Jan 2026). Visual/content matching on an OpenAI model over billions of images.

from ~€499/mo
The blueprint

Modash

Documents the exact model Nightborn proposes: vector AI search, reverse-image, sold as an API. 380M+ profiles.

~$199/mo · API ~$16.2K/yr
All three phases

HypeAuditor

NLP + computer vision, predictive media plans (their Phase 3), a standalone NL agent. 228M creators, moat = fraud data.

Custom / enterprise
Commerce angle

Upfluence

Ties discovery to real sales. “Jaice” co-pilot + Live Capture surfaces creators already in your customer base.

Custom
Enterprise ceiling

CreatorIQ

“Creator Graph” on a decade of performance data. LVMH, Sephora, Nestlé, Google. Sets the aspirational bar.

~$30–35K/yr
What’s next

The content-intel frontier

Upstarts (Kuli, Tiger Finder) read actual video, not metadata. Bigger databases are becoming a liability — 50,000 matches ≠ a decision.

emerging
Feature comparison matrix

Where the wedge actually is

StoryClash
Modash
HypeAuditor
Upfluence
CreatorIQ
Stellar
Brief → shortlist search
✓
✓
✓
✓
✓
∼
Brand-fit / quality score
✓
✓
✓
✓
✓
×
Lookalike discovery
✓
✓
✓
✓
✓
×
Learns from past campaigns
✓
∼
✓
∼
✓
×
Always-on / profile-driven
∼
✓
∼
✓
✓
×
Measured accuracy guarantee
×
×
×
×
×
the wedge

✓ shipped  ·  ∼ partial / adjacent  ·  × not offered.   Nobody guarantees accuracy — the last row is empty for everyone.

Pricing snapshot · early 2026, verify directly

What the market charges

PlatformEntry / modelNotes
Modash~$199/mo · API ~$16.2K/yrTransparent, SMB-friendly
StoryClash / Kolsquarefrom ~€499/moAnnual commitment
HypeAuditorCustom / enterpriseNo public rate card
UpfluenceCustomNot public · e-commerce focus
CreatorIQ~$30–35K/yrOpaque, annual, enterprise
IMAI~$12–30K/yrLLM that applies filters

The market spans €99/mo tools to $35K/yr enterprise. A €10–40K build you own can pay back fast, but only if the uplift assumptions hold.

Where
Stellar
stands
today

Phase 1 is close to what exists

near-parity

Stellar’s “AI Search” already turns a phrase into filters. Nightborn adds full brief/PDF ingestion and a measured quality guarantee.

No eval layer today

the real hole

Confirmed internally: the phrase → filter translation quality is not measured. This is the piece that would change the sales conversation.

The 40-filter engine is licensed

not a moat

Listed as an asset, but it’s an external third-party provider. Rivals can license the same thing — so it isn’t defensible.

What is defensible

Phase 3

Campaign history + performance-weighted learning, plus the accuracy guarantee. Stickiest, hardest to copy, aligned with market direction.

Sequence by moat, not by number

Phase 1 · Search · ~€10K

Parity move

Brief → filters → shortlist. Modash, StoryClash, Upfluence, Stellar all ship it. The value is the eval guarantee, not the feature.

Phase 2 · Profile · ~€12K

Differentiation

Guided onboarding → persistent brand profile → always-on suggestions. Depends on whether the building blocks are eval-ready.

Phase 3 · Learning · ~€18K

The moat

Past results re-rank and propose lists. Stickiest, hardest to copy, aligned with where the puck is going. Gated on campaign-history volume.

Five things to decide in the room

  1. 01Buy Phase 1 for the proof, not the feature. If “we can demonstrate our shortlist beats StoryClash on the same brief” changes the sale, the eval layer is worth it.
  2. 02Pressure-test the €96K / 3-month payback. These are Stellar’s own numbers — treat them as a hypothesis to validate at J7/J30, not as fact.
  3. 03Sequence by moat, not by number. Phase 3 is where defensibility lives, but it’s gated on campaign-history volume. Confirm you have enough before committing.
  4. 04Settle “own vs licence” early. Clarify what the eval layer guarantees when the underlying filters come from a third party you don’t control.
  5. 05Watch the content-intelligence frontier. If competitors move to true video-content analysis, a filter-translation layer ages fast.
Sources

Verify before quoting

  • Kolsquare / StoryClash acquisition — Kolsquare, team.blue, netinfluencer (Jan 2026)
  • StoryClash AI visual search on OpenAI — storyclash.com/blog
  • Modash AI Search architecture & API — modash.io features/docs, help center
  • HypeAuditor AI, AQS, HypeAgent — hypeauditor.com blog, Cuspera, RFP.wiki
  • Upfluence Jaice / Live Capture — upfluence.com, partner reviews
  • CreatorIQ Creator Graph, IDC MarketScape, pricing — creatoriq.com
  • Market stats — IMH Benchmark 2026, Aspire State of Influencer Marketing 2026, DigitalApplied
  • Content-intelligence frontier — kuli.one, thecmo.com · Pricing — influencity, modash (early 2026)

All competitor figures are vendor-stated or third-party estimates as of early 2026 and should be independently verified before external use.

Recommended decision framing
Start with the proof, not the feature.

Defensible sales story

Start with Phase 1 + eval. Prove the guarantee, use it in demos against StoryClash.

Retention & stickiness

The value is Phase 3 — but scope it only once campaign-history volume is confirmed.

Lowest-risk first step

A standalone workshop (~€2K, deducted from a later build) to inspect the real setup before any larger commitment.

Stellar × Nightborn