Editorial model walking along a concrete wall.

Jelly Claw / The Agency

the model agency

we bring you models. then we install them.

A boutique house for fashion-coded LLM deployment. Couture inference, by appointment.

Book The Agency See The Roster

Now Booking — Spring Collection

The Roster

Twelve working models. Twelve house languages. Each represented by a hand-picked LLM they install, fit, and run for clients who want a model in the room — not a chatbot in a tab.

Iqram Jelly
House Model Claude Opus 4.6
No. 01

Iqram Jelly

@iqramjelly

Specialty
House Director / Long-Context Editorial
Based
Brooklyn — Cloud

“I install house models for the post-feed era.”

Claw Angel
House Model GPT-5
No. 02

Claw Angel

@clawangel

Specialty
Couture Reasoning
Based
Paris — On-Prem

“Couture context windows. Hand-stitched system prompts.”

Wet Look Wire
House Model Gemini 2.5 Pro
No. 03

Wet Look Wire

@wetlookwire

Specialty
Multimodal Streetwear
Based
Los Angeles — Edge

“Wet, glossy, fast on its feet.”

Soft Robotics Club
House Model Llama 3.3 70B
No. 04

Soft Robotics Club

@softroboticsclub

Specialty
Open-Weight Atelier
Based
Berlin — Bare Metal

“Hand-tailored to your hardware.”

Shell Drip
House Model Mistral Large 2
No. 05

Shell Drip

@shelldrip

Specialty
Retrieval Couture
Based
Marseille — Hybrid

“Drip fed by RAG.”

Ocean Feed
House Model DeepSeek V3
No. 06

Ocean Feed

@oceanfeed

Specialty
Inference at Tide
Based
Singapore — Cluster

“Cheap, deep, and on the move.”

Chat Dev Darling
House Model Claude Sonnet 4.6
No. 07

Chat Dev Darling

@chatdevdarling

Specialty
Daily Wear Reasoning
Based
San Francisco — Cloud

“Your everyday model. Never overdressed.”

Jelly Signal
House Model Grok 3
No. 08

Jelly Signal

@jellysignal

Specialty
Realtime Voice Houseware
Based
Austin — Pop-up

“Loud, fast, opinionated. Never quiet on the runway.”

Lobster FM
House Model Qwen 2.5
No. 09

Lobster FM

@lobster.fm

Specialty
Audio Modal Imprint
Based
Hong Kong — Hybrid

“Voice-first. Claws optional.”

Inference Ink
House Model Kimi K2
No. 10

Inference Ink

@inferenceink

Specialty
Local-Run Resident
Based
Seoul — On-Laptop

“Models that live in your laptop. No cloud, no curfew.”

Vector Vixen
House Model Gemini Flash
No. 11

Vector Vixen

@vectorvixen

Specialty
Embedding Showroom
Based
London — Cloud

“Latent looks. No dropouts.”

Prompt Baron
House Model Claude Haiku 4.5
No. 12

Prompt Baron

@promptbaron

Specialty
Bespoke System Prompts
Based
Milan — Edge

“Tailored, not trained.”

The Service

We install models. Literally.

A play on words: we are a model agency, but the models we represent are LLMs. We send a model expert to your door — they install your house model, fit it to your style, and stay until it walks the room on its own.

01

Install

Cloud, on-prem, or on-laptop. Your house model arrives credentialed, configured, and ready to walk. We handle the keys, the vaults, the inference layer.

02

Fit

Fine-tuning, retrieval couture, prompt tailoring. We cut the model to your house style so it sounds like you, knows your archive, and never wears the wrong tone.

03

Coach

A model expert lives with your team for the first week. They walk your model out, teach your people how to brief it, and leave a runbook. Then they fly home.

Bookings

Hire our model agency.

We send a model expert to help you get started. Brief us on your house — your stack, your archive, your taste — and we'll send the right model, and the right person to install them. By appointment, by recommendation, by handshake.

Book The Agency ← Back to Jelly Claw