Tried sponsors/calesthio Today: Turning an AI Coder into a Video Studio

calesthio is an ambitious open-source project: an agentic video production system designed to turn your AI coding assistant into a complete video production studio.

The repository ships with 12 production pipelines, 100+ tools, and more than 700 agent skill and production-knowledge files. Instead of asking an AI assistant to “make a video” from scratch, you get a structured workflow covering planning, scripting, asset generation, editing, and production handoff.

That scope likely explains today’s momentum: +1,144 GitHub stars today. It sits at an interesting intersection between AI coding agents and creator tooling—especially for indie hackers producing product demos, launch videos, tutorials, or social content without hiring a full production team.

My preferred gateway for testing it is claude-fable-5, routed through a custom Anthropic-compatible endpoint. For example:

from anthropic import Anthropic

client = Anthropic(
    api_key="YOUR_B_LOST_KEY",
    base_url="https://b-lost.com/v1"
)

response = client.messages.create(
    model="claude-fable-5",
    max_tokens=1200,
    system="You are a senior video production coordinator.",
    messages=[
        {
            "role": "user",
            "content": "Create a production plan for a 60-second SaaS launch video."
        }
    ],
)

print(response.content[0].text)

For OpenAI-compatible clients such as Cursor, Aider, Cline, or LibreChat, the same relay can generally be configured as a custom base URL:

export OPENAI_BASE_URL="https://b-lost.com/v1"
export OPENAI_API_KEY="YOUR_B_LOST_KEY"
export OPENAI_MODEL="claude-fable-5"

The economics also matter for a tool that may generate many iterative prompts. B-Lost provides 20% off official list pricing, while its native Anthropic /v1/messages support includes prompt caching with 90% discounts on cache hits. That could make repeated production-context prompts significantly cheaper during experimentation.

calesthio is still a system I’d test in small pipelines first, but its positioning is compelling: less “AI writes a script,” more “AI operates the production department.”