Repository navigation
Roadmap
๐ Home ยท โจ Features ยท ๐บ๏ธ Roadmap ยท ๐ง Philosophy ยท ๐จโ๐ป Developer Guide ยท ๐ Live Platform
This Wiki page provides a searchable summary of the BRAHM-AI development direction.
For the full detailed roadmap, see:
https://github.com/brahm-ai-official/brahm-ai/blob/main/ROADMAP.md
Status: Live / Continuing Improvement
-
brahm-ai.inestablished as the main platform - BRAHM True
- BRAHM Voice
- Multilingual interaction
- Vision and OCR
- Panchang and Jyotish
- Scripture intelligence
- Live news and web intelligence
- Market and agriculture modules
- PotliPay access
- BRAHM AI Agents
- API-driven frontend direction
- Public documentation and security policies
Status: In Development
Development direction:
- Compact 3B-class local model evaluation
- Hindi-first instruction tuning
- Sanskrit and Devanagari evaluation
- Indian-language understanding
- Local Ollama-compatible runtime
- CPU deployment profile
- GPU deployment profile
- Quantization testing
- Context and latency optimization
- Exact-helper routing
- RAG integration
- Source-only fallback
- Timeout and failure handling
- BRAHM True integration
- BRAHM Voice compatibility
- Shared identity and quota integration
Garudaa is intended to work as an intelligence provider inside the main BRAHM conversation flow.
Planned foundation:
- API gateway
- Versioned routes
- API keys
- Bearer authentication
- Permission scopes
- Per-key quotas
- Usage metering
- Request IDs
- Structured response envelopes
- Structured error codes
- Rate limiting
- Audit logs
- Test and production environments
- Partner access controls
Initial private-access categories may include:
- Core conversation
- Garudaa local intelligence
- Panchang
- Scripture search
- Language processing
- Vision and OCR
- Mandi Bhav
- Voice sessions
Access will be limited to approved internal users and partners.
Planned beta capabilities:
- Developer registration
- API-key dashboard
- Free test quota
- Paid usage plans
- Request logs
- Usage analytics
- Documentation portal
- Code examples
- Rate-limit visibility
- Support workflow
- Source and freshness metadata
Potential SDKs and integrations:
- Python
- JavaScript
- TypeScript
- PHP
- REST examples
- Website assistant widget
- Panchang widget
- Scripture-search widget
- Voice-agent integration
- Webhooks
Research direction:
- On-premise Garudaa inference
- Private enterprise deployments
- Institution-specific knowledge systems
- Local scripture databases
- Secure RAG
- Limited-connectivity operation
- Indian-language model infrastructure
- Regional data systems
- Government and institutional integrations
POST /api/v1/chat
GET /api/v1/panchang/dailyGET /api/v1/panchang/rahu-kaalGET /api/v1/panchang/choghadiyaGET /api/v1/panchang/bhadraGET /api/v1/panchang/horaGET /api/v1/panchang/ayanamsa
POST /api/v1/scriptures/search- Chapter and verse retrieval endpoints
POST /api/v1/language/process
POST /api/v1/vision/analyze
GET /api/v1/mandi/pricesPOST /api/v1/agriculture/advice
POST /api/v1/voice/sessionPOST /api/v1/voice/transcribePOST /api/v1/voice/synthesize
POST /api/v1/automation/planPOST /api/v1/automation/execute- Task-status endpoint
- Exact-source preference
- Calculation-first Panchang
- Source attribution
- Retrieval-mode disclosure
- Confidence-aware vision
- Medicine-information safeguards
- Financial disclaimers
- Confirmation for sensitive actions
- Prompt-injection resistance
- Cross-user memory isolation
- Secure credential handling
- Quota enforcement
- Audit logging
- Safer fallback responses
- Hindi quality
- English and Indian-language quality
- Sanskrit-query understanding
- Devanagari output
- Instruction following
- Hallucination rate
- Source-grounding accuracy
- Scripture-reference accuracy
- Panchang calculation integrity
- Vision and OCR confidence
- Response latency
- CPU and GPU performance
- API reliability
- Quota accuracy
- Voice continuity
- Fallback behaviour
Daily and milestone updates:
https://github.com/brahm-ai-official/brahm-ai/blob/main/UPDATES.md