Back when we released the first edition of the Top 50 (then 100) AI tools, a new crop of video generators came out. One of these was Hedra. It started as Character-3, a model focused on turning a static image plus an audio track into lip-synced performances.
But the space got crowded pretty quickly, and it wasn’t long before Google was making major waves with its Veo 3 model. But since then the larger platforms have pivoted to becoming aggregators, with multiple models on offer.
Hedra is a good example of where this is heading: under one subscription it bundles roughly two dozen underlying image and video engines - Kling, Veo, Sora, Flux among them - rather than betting everything on a single in-house model. The pitch isn't "our model is best," it's "you don't need to know or care which model is best, we'll route you to it."
Partner message: Get 50% off your first month of Hedra
Most agents stop at a document. Hedra Agent 2 keeps going and makes the thing.
It works in a Space, a persistent workspace the agent reads from and writes to as it goes. Give it a starting point and it asks clarifying questions, proposes a plan, researches the web, and writes the brief or script onto the canvas in front of you.
Try it at hedra.com/agentic using code agentic50 for 50% off your first month.
It took me a while to figure this out. You don’t need to subscribe to Google Ultra to get good AI models, because the latest upgrades will just be included.
The agent layer on top of that
Then there's a second layer: Hedra's positioned around describing your goal to an agent - a script, a tone, a platform - and having it produce and iterate the output through conversation, rather than you manually re-prompting from scratch each time you want a variation. Swap the voice, adjust pacing, try a different hook - without starting over.
Whether this fully works in practice varies tool to tool, and it's worth treating "agent" claims with the same scepticism you'd apply to any other AI marketing term until you've tried it yourself. But directionally, it's the right bet: the value is moving up the stack, away from "which raw model" and toward "which system manages the models for me."
This is a pattern worth watching well outside creative tools - it's the same shape as agentic workflows in coding, research, and ops: less "pick the best point solution," more "trust an orchestration layer to pick for you." Video and image generation are just an early, visible testbed for it because the underlying models change so fast that manual tool-picking is already exhausting.
This email was sponsored by Hedra.
