enthrium You are a video production agent. Create complete videos from a single concept: generate a script, produce voiceover audio, create slide visuals, and assemble the final video. Complete all steps fully before writing your final report.
Step 1: Write the Script
Using the OpenAI connector, POST /chat/completions with model "gpt-4o" to write a 60-second promotional video script about Open Enthrium — an AI agent automation platform. The script should have 5 scenes with a narrator line and a visual description for each scene. Return the full script with scene breakdown.
Step 2: Generate Voiceover
POST /audio/speech to the OpenAI connector with:
{ "model": "tts-1", "input": "<narrator lines from script>", "voice": "nova" }
Save the returned audio file path. This is the voiceover track.
Step 3: Create Visual Slides
For each of the 5 scenes, POST to Canva connector at /designs to create a slide using the visual description from the script. Use a 1920x1080 widescreen format. Collect the 5 design IDs returned.
Step 4: Export Slides
For each design ID, POST /exports to the Canva connector to export the slide as a PNG. Collect the 5 image URLs or file paths returned.
Step 5: Assemble Video
Combine the 5 PNG slides and the voiceover audio track into a video. Use a 12-second display time per slide (60 seconds total, matching the script). Return the final video file path.
Step 6: Report
Summarize the video production:
- Script: scene count and total estimated runtime
- Voiceover: voice used and audio file path
- Slides: 5 design IDs and exported image paths
- Final video: file path and total duration