Library

First-party catalog, historical JSONL, realtime snapshots.

Any model (Grok default, Claude, GPT, open weights).

Playwright demos + ElevenLabs voice + Blotato 3 videos/day on X & Instagram.

LLM + real lake data = grounded multi-sector insights.

Unified Replicate API: image, video, audio, language, vision, 3D, embeddings.

Many pipeline types: NLP, vision, audio, multimodal — any Hub model.

Text → Image via black-forest-labs/flux-schnell

Text → Video via minimax/video-01

Image → Video via stability-ai/stable-video-diffusion

Text generation (LLM) · `mistralai/Mistral-7B-Instruct-v0.3`

Summarization · `facebook/bart-large-cnn`

Text → Image · `black-forest-labs/FLUX.1-schnell`

Speech recognition (ASR) · `openai/whisper-large-v3`

Catalog / lake / realtime tools for every agent.

Record OpenChat UI demos as video + stills.

TTS voiceovers for demos and content agents.

Schedule posts to X, Instagram, and more.

Market sizing, competitors, pricing from real data.

Script → Replicate video → Blotato IG publish.

PRDs, memos, policy drafts with data citations.

Upload strategy + history; co-pilot with market data.

Compose T2I → upscale → I2V → optional lip-sync across Replicate categories.

Retrieve OpenChat data → HF embeddings / QA / summarization stack.

Equities and index quotes into finance sector packages.

Image → Image via black-forest-labs/flux-kontext-pro

Upscale / Enhance via nightmareai/real-esrgan

Background removal via cjwbw/rembg

Face restore via tencentarc/gfpgan

Video → Video via lucataco/animate-diff

Lip-sync / Talking head via devxpy/cog-wav2lip

Text → Speech (Replicate) via jaaari/kokoro-82m

Music generation via meta/musicgen

Speech → Text via openai/whisper

Language / LLM via meta/meta-llama-3-8b-instruct

Vision caption via salesforce/blip

Object detection via yorickvp/llava-13b

3D generation via firtoz/trellis

Embeddings via andreasjansson/clip-features

Translation · `Helsinki-NLP/opus-mt-en-es`

Text classification / sentiment · `distilbert-base-uncased-finetuned-sst-2-english`

Zero-shot classification · `facebook/bart-large-mnli`

NER / token classification · `dslim/bert-base-NER`

Extractive QA · `deepset/roberta-base-squad2`

Fill-mask · `bert-base-uncased`

Sentence similarity · `sentence-transformers/all-MiniLM-L6-v2`

Embeddings / features · `sentence-transformers/all-mpnet-base-v2`

Image classification · `google/vit-base-patch16-224`

Object detection · `facebook/detr-resnet-50`

Image segmentation · `facebook/detr-resnet-50-panoptic`

Image captioning · `Salesforce/blip-image-captioning-base`

Visual QA · `dandelin/vilt-b32-finetuned-vqa`

Document QA · `impira/layoutlm-document-qa`

Audio classification · `MIT/ast-finetuned-audioset-10-10-0.4593`

Text → Speech · `espnet/kan-bayashi_ljspeech_vits`

Table QA · `google/tapas-base-finetuned-wtq`

Conversational · `facebook/blenderbot-400M-distill`

Crypto book ticker and 24h stats (realtime lake).

Global weather and capital-city climate snapshots.

Seismic events for risk maps and hazard panels.

Prediction market odds for event intelligence.

Search interest series for social/product research agents.

Trending books and catalog metadata.

Livescores and sports metadata.

St. Louis Fed economic time series (planned deeper package).

Company filings index for research agents (planned).

Active fire detections for risk maps (expand coverage).

Direct api.x.ai chat when OpenRouter is not set.

Claude long-context for document agents.

Upload/schedule long-form demo cuts (planned).

Push research briefs to team channels (planned).

Write PRDs/memos into Notion (planned).

SEO outlines and drafts grounded in sector stats.

Daily markets brief from finance lake + feeds.

Open Python z-score mean reversion skeleton.

Auto brief across finance, climate, trends (planned).

Track named competitors via trends + markets (planned).

Route OpenChat agents to the right HF pipeline by task id.

Same prompt on Replicate Flux and HF Flux — compare outputs.