> ## Documentation Index
> Fetch the complete documentation index at: https://docs.forii.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Roadmap

> What's coming to Forii

## Coming Soon

India-differentiated capabilities and essential observability.

| Feature                     | Description                                                         |
| --------------------------- | ------------------------------------------------------------------- |
| **Paid Plans**              | Starter, Pro, Enterprise tiers with higher RPM/TPM limits           |
| **Credit System**           | Prepaid credits (1 = ₹1), UPI/card/netbanking via Razorpay          |
| **GST Invoices**            | CGST/SGST and IGST-compliant PDF invoices with your GSTIN           |
| **Speech-to-Text**          | `forii/saarika-v2` — 22 Indic languages, 8kHz telephony audio       |
| **Text-to-Speech**          | `forii/bulbul-v2` — 30+ voices across Hindi, Tamil, Telugu, Bengali |
| **Vision / Multimodal**     | Image understanding — Aadhaar, PAN, invoice OCR via `image_url`     |
| **Reranking**               | Cross-encoder reranking for RAG — Indic-optimized                   |
| **Batch Inference**         | Asynchronous large-scale processing at 50% lower cost               |
| **Fine-Tuning (SFT)**       | Supervised fine-tuning on Indic-language data                       |
| **Response headers**        | TTFT, region, request-id, cached-tokens headers                     |
| **CLI tool**                | `forii chat --verbose`, `forii usage`, `forii models list`          |
| **Usage API**               | Detailed usage breakdown by model, key, date                        |
| **Request annotations**     | Cost attribution by team/project/environment                        |
| **Token-level rate limits** | TPM per key per model                                               |

## Planned

Platform features, advanced inference, and enterprise tooling.

| Feature                     | Description                                                 |
| --------------------------- | ----------------------------------------------------------- |
| **Deployments**             | Dedicated GPU with autoscaling and scale-to-zero            |
| **Prompt caching**          | Automatic prefix caching with discount on cached tokens     |
| **Anthropic compatibility** | `/inference/v1/messages` endpoint for Claude SDK users      |
| **Responses API**           | Stateful conversations with `previous_response_id`          |
| **Legacy completions**      | `/inference/v1/completions` — text-in/text-out              |
| **DPO fine-tuning**         | Direct Preference Optimization on preference pairs          |
| **RFT**                     | Reinforcement Fine-Tuning with custom evaluators            |
| **Advanced dashboard**      | Latency percentiles, error trends, alerts, region breakdown |
| **Prometheus metrics**      | Metrics endpoint for Grafana, Datadog, OTel                 |
| **External integrations**   | LangSmith, Langfuse, Grafana Cloud, Datadog                 |

## Stay updated

Watch [GitHub](https://github.com/forii-in) or join our [Discord](https://discord.gg/forii) for release announcements.