Introducing Rendyr v1 — autonomous video generation for crypto protocols. Get early access →
Rendyr deploys AI agents that scan your dApp, write a script, navigate with a real wallet, record the session, and deliver a publish-ready video ad. Fully autonomous. Minutes, not weeks.
|
AUTONOMOUS PIPELINE
Rendyr orchestrates specialized AI agents that work like a production crew. Scanner reads your protocol. Writer crafts the narrative. Navigator drives the browser. Producer delivers the final cut. All autonomous.
Point Rendyr at any dApp URL. The scanner agent reads the page using vision models, identifies the protocol type, extracts market data, and builds a structured profile. No CSS selectors. No scraping. Pure visual intelligence.
Vision Scan
Sees the page like a human.
Auto-Classify
DEX, prediction market, lending, NFT.
Data Extraction
Markets, odds, TVL, APY.
Chain Detection
Polygon, Base, Gnosis, Neura.
# Scan a protocol
$ rendyr scan \
--url "https://roar.fund/market/eth-5k" \
--chain polygon
# Output: structured protocol profile
{
"type": "prediction_market",
"markets": 24,
"tvl": "$1.2M",
"chain": "polygon"
}A dedicated LLM agent analyzes the protocol profile, your target audience, and selected tone to write original narration. Not fill-in-the-blank templates. Every script is unique, with a hook, narrative arc, and call to action calibrated to your audience.
Custom Hooks
Attention-grabbing openers.
Audience Targeting
Crypto traders, DeFi degens, normies.
Tone Control
Confident, educational, hype.
Quality Gate
AI reviewer blocks weak scripts.
# Generate a script
$ rendyr script \
--protocol "roar-fund" \
--tone "confident, crypto-native" \
--audience "crypto traders" \
--format vertical
# Output: timestamped narration
{
"hook": "This prediction market is paying...",
"scenes": 6,
"duration": "28s",
"caption": "How to bet on ETH hitting 5k..."
}Autonomous browser agents launch a real Chromium instance with MetaMask, navigate to your dApp, connect a wallet, and execute real on-chain interactions. The entire session is recorded with cinematic cursor movement and pacing. Not mockups. Real transactions.
Real Browser
Playwright + Chromium + MetaMask.
Vision Navigation
Agents see the UI, not selectors.
Wallet Actions
Connect, sign, bet, swap.
Cinematic Capture
Cursor highlights, zoom, pacing.
# Record a session
$ rendyr record \
--protocol "roar-fund" \
--script "script_abc123" \
--wallet metamask \
--format 9:16
# Status: recording in progress
> Connecting wallet... done
> Navigating to market... done
> Placing bet... done
> Recording complete. 28.4sThe final agent composites everything: recording footage, AI voiceover synced to the action, cursor spotlights, lower thirds, transitions, and background music. Output is a publish-ready video optimized for TikTok, X, or YouTube Shorts.
AI Voiceover
Synced narration from script.
Motion Graphics
Lower thirds, transitions, bumpers.
Format Options
Vertical 9:16 or horizontal 16:9.
Platform Ready
TikTok, X, YouTube Shorts.
# Produce final video
$ rendyr produce \
--recording "rec_xyz789" \
--voice "confident-male" \
--music ambient \
--export mp4
# Output
{
"video_url": "https://cdn.rendyr.xyz/v/...",
"duration": "28s",
"format": "1080x1920",
"size": "12.4MB"
}THE MOAT
Competitors hardcode CSS selectors that break when a site updates. Rendyr uses multimodal vision models that adapt to any UI, any layout, any chain. Zero maintenance.
Not one monolithic prompt. Rendyr coordinates specialized agents: scanner, scriptwriter, navigator, recorder, quality gate. Each does one thing extremely well.
Every video feeds back into the system. The quality gate evaluates output, flags weak footage, and the pipeline learns which patterns produce the best content.
Rendyr builds a fingerprint of every protocol it encounters. More dApps processed means faster, more accurate handling of new ones. Network effects compound.
Connecting a real wallet, switching chains, and executing live transactions inside automated video is an extremely hard engineering problem. Solved.
Every recording teaches new UI patterns. Every script teaches new angles. Every failure teaches what to avoid. The longer Rendyr runs, the wider the gap.
FLYWHEEL EFFECT
Rendyr is not a static tool. It is a learning system. Every dApp it navigates teaches the vision model new UI patterns. Every script sharpens the narrative engine. Every failed recording teaches the quality gate what to catch.
After processing hundreds of protocols, Rendyr recognizes common layouts instantly, knows which hooks perform best, and navigates unfamiliar dApps on the first try. That is not something you shortcut with more compute.
PRICING
3 videos per month
15 videos per month
50 videos per month
Stop paying freelancers $500 per video. Let AI agents produce publish-ready content in minutes, not weeks.