The agent calls asticaVision's OCR endpoint with the image URL or Base64 data. The API returns extracted text with word-level bounding boxes. Model version 2.0_full or higher is required for OCR support. Results include position coordinates for each detected word.

Yes. Astica AI's Analyze Audio accepts WAV and MP3 files via HTTPS URL or Base64 encoding. The speech-to-text model processes the audio and returns full transcription. Streaming mode provides partial results for longer files.

Tars requires your Astica API key (tkn parameter). Generate this from your Astica AI account. The key authenticates requests to vision and audio endpoints. Usage is billed per request based on Astica's compute credit system.

No. Tars sends media to Astica's API for processing and receives only the extracted text or transcription. Original images and audio files are not stored by Tars. All media processing occurs in Astica's infrastructure.

Yes. Astica's OCR capabilities include handwritten text recognition. The text_read parameter returns detected handwriting along with printed text. Accuracy varies based on handwriting legibility and image quality.

The agent processes available content and returns what can be extracted. For images, blurry or low-resolution sections may have reduced accuracy. For audio, background noise affects transcription quality. The agent reports confidence levels when available.

Standard uploads just store files. Astica AI integration extracts actionable information from media. Your agent understands image content and spoken words, enabling automated workflows based on document data and voice requests.

Yes. The agent can call the OCR endpoint for each image sequentially. Each image is processed independently and results can be combined. For bulk processing, consider batch workflows with consolidated results.