agentleFS
Sign inStart free
ankitjh4ankitjh4/indic-ai-skills/skills/sarvam-ai/SKILL.md

sarvam-ai

Skill65 starsChanged 7 months agoMITView on GitHub

Indian AI toolkit powered by Sarvam AI — text-to-speech, speech-to-text, document intelligence, translation, transliteration, language detection, and chat completion across 23 Indian languages. Use when working with Indian languages, Hindi/Tamil/Bengali text, Sarvam AI, or when the user needs translation, transcription, or TTS for South Asian languages.

  • Reads credentials

1 other project uses a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

1 rule. About 0 are common, 1 is distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Batch workflow: create job → upload files → start → poll status (Accepted → Pending → Running → Completed) → download results.

    line 58

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Sarvam AI — Indian Language Toolkit

Comprehensive AI toolkit for 23 Indian languages: TTS, STT, Document Intelligence, Translation, Transliteration, Language Detection, and Chat.

Setup

  1. Get a free API key at https://dashboard.sarvam.ai
  2. Set environment variable: export SARVAM_API_KEY="your-api-key"

Supported Languages

hi-IN Hindi, en-IN English, bn-IN Bengali, gu-IN Gujarati, kn-IN Kannada, ml-IN Malayalam, mr-IN Marathi, or-IN/od-IN Odia, pa-IN Punjabi, ta-IN Tamil, te-IN Telugu, ur-IN Urdu, as-IN Assamese, bodo-IN/brx-IN Bodo, doi-IN Dogri, ks-IN Kashmiri, kok-IN Konkani, mai-IN Maithili, mni-IN Manipuri, ne-IN Nepali, sa-IN Sanskrit, sat-IN Santali, sd-IN Sindhi

Text-to-Speech

python3 scripts/tts.py "नमस्ते, आप कैसे हैं?" --language hi-IN --speaker meera
ParameterDefaultDescription
text—Text to convert (max 2500 chars)
--languagehi-INLanguage code
--speakermeeraVoice name
--outputoutput.wavOutput file
--sample-rate24000Audio sample rate

Speakers — Female: Meera, Priya, Neha, Simran, Kavya, Ishita, Shreya, and more. Male: Shubh, Aditya, Rahul, Amit, Dev, Arjun, and more.

Speech-to-Text

Three modes: REST (quick, <30s), WebSocket (real-time streaming), Batch (long audio, diarization).

# REST — quick transcription
python3 scripts/speech_to_text.py rest audio.mp3

# WebSocket — real-time streaming
python3 scripts/speech_to_text.py websocket audio.wav

# Batch — multiple files with speaker diarization
python3 scripts/speech_to_text.py batch audio1.mp3 audio2.mp3 --diarization --num-speakers 3 --output-dir ./transcripts/

Batch workflow: create job → upload files → start → poll status (Accepted → Pending → Running → Completed) → download results.

Formats: WAV, MP3, AAC, AIFF, OGG, OPUS, FLAC, MP4/M4A, AMR, WMA, WebM, PCM

Document Intelligence

Extract text from PDFs and images (JPEG/PNG).

python3 scripts/document_intelligence.py document.pdf --language hi-IN --format md
python3 scripts/document_intelligence.py --job-id <id> --download -o ./output/

Formats: md (default), html, json. Max 200 MB, 500 pages.

Translation

# Auto-detect source, translate to Hindi
python3 scripts/text_processing.py translate "Hello, how are you?" --target hi-IN

# Mayura model with colloquial mode
python3 scripts/text_processing.py translate "What's up?" --target hi-IN --model mayura:v1 --mode modern-colloquial

Models: sarvam-translate:v1 (23 languages), mayura:v1 (12 languages, supports modes and transliteration)

Modes (mayura only): formal, modern-colloquial, classic-colloquial, code-mixed

Transliteration

python3 scripts/text_processing.py transliterate "नमस्ते" --source hi-IN --target en-IN
python3 scripts/text_processing.py transliterate "namaste" --source en-IN --target hi-IN --spoken-form

Language Detection

python3 scripts/text_processing.py detect "नमस्ते दुনিয়া"
# Output: Language: hi-IN, Script: Deva

Chat Completion

Two models: sarvam-105b (flagship, complex reasoning) and sarvam-m (efficient, general chat).

python3 scripts/text_processing.py chat "Explain quantum computing" --model sarvam-105b
python3 scripts/text_processing.py chat "What is the capital of India?" --model sarvam-m --temperature 0.8

Resources

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

Check your setup, free