TIR Model Endpoints

Ship models as APIs. Scale to zero when quiet.

Pick a model from the catalog or your Hugging Face account, choose a GPU, and get an authenticated OpenAI-style endpoint served from Indian data centers — with autoscaling you actually control.

6
Runtimes: vLLM · SGLang · Triton · PyTorch · Dynamo · Custom
0
Idle replicas with scale-to-zero
₹49/hr*
GPU endpoints start
99.9%
Uptime SLA
DeepSeek R1DeepSeek V3Llama 4 MaverickLlama 4 ScoutGPT-OSS 120BHermes 3 405BDolphin Mixtral 8x22BSarvam-2B

*Indicative — see the live rate card for current pricing.

Inference infrastructure, minus the infrastructure

Every feature below is live in the TIR console today.

One-click catalog deploys

Serve state-of-the-art open models from the TIR catalog, or hit the shared GenAI API (GPT-OSS 120B, Llama 3.1 8B, Qwen3 32B, Sarvam and more) without deploying anything yourself.

Hugging Face & Model Repository

Point an endpoint at any HF model or a TIR Model Repository on E2E Object Storage — files download to the container automatically at start. Validate model and config before deploy.

Custom containers

Bring your own Docker image from a public or private registry, optionally attached to a TIR Model for automatic weight downloads. Full control over the serving stack.

Metric-driven autoscaling

Scale on concurrent requests, requests/sec, or runtime metrics like vllm:num_requests_running, sglang:gen_throughput and Triton queue depth. Tune idle timeout and cooldown to balance latency vs cost.

Async & WebSocket APIs

Queue long-running jobs with async invocation — get a request_id, results land as JSON in your EOS dataset, workers bill only while the queue is non-empty. Streaming? Connect over wss://.

Private, observable serving

Security Groups and VPC attach keep endpoints off the public internet. Centralized inference and per-replica logs, hardware + service metrics, access logs and email alerts come standard.

From model to production API in four steps

The actual deploy flow in the TIR console.

Step 1

Pick a source

Catalog model, Hugging Face repo, TIR Model Repository, or your own container image.

Step 2

Choose runtime & GPU

vLLM, SGLang, Triton or PyTorch; select GPU plan and replicas. Tune engine args — dtype, quantization, context length — if you need to.

Step 3

Set scaling

Active Workers, Max Workers, metric type and target value. Different min/max = autoscaling on. Zero minimum = scale-to-zero.

Step 4

Call it

Create an API token, copy the sample cURL/Python from the console, and test in the built-in playground before wiring up your app.

OpenAI-compatible out of the box

Every endpoint gets a root URL; append standard routes like /v1/chat/completions.

# illustrative — copy the exact URL from your endpoint's API Request tab
curl https://infer.e2enetworks.net/project/p-XXXX/endpoint/is-XXXX/v1/chat/completions \
  -H "Authorization: Bearer $TIR_API_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "deepseek-r1",
    "messages": [{"role": "user", "content": "Summarise this contract clause:"}],
    "max_tokens": 400
  }'

Tokens live under API Tokens in your TIR project. The playground supports chat models with temperature, top-p, top-k and repetition-penalty controls.

Autoscaling signals, compared

Choose the metric that matches your traffic shape.

Metric typeScales onBest forExample target
Concurrent Request CountIn-flight requests per workerLatency-sensitive LLM chat, where each request holds the GPU for the full generation10 concurrent requests
Request Rate per SecondRequests/sec arriving per workerHigh-throughput, short-lived calls — embeddings, classification50 req/s
Custom (vLLM)vllm:num_requests_running, vllm:gpu_cache_usage_perc…Preventing KV-cache OOM, queue-depth-aware scaling32 running requests
Custom (SGLang)sglang:num_queue_reqs, sglang:gen_throughput…Token-throughput-driven servingQueue > 8
Tritonnv_inference_pending_request_count, nv_gpu_utilization…Multi-model serversGPU util 80%
TorchServeGPUUtilization, ts_queue_latency_microseconds…Classic PyTorch servingQueue latency

Idle Timeout keeps warm workers available (billed); Initial Cooldown Period prevents flapping before scale-to-zero.

Fine-tune first, then serve it here

Models trained or fine-tuned on TIR deploy to endpoints in a few clicks — same project, same storage.

Frequently Asked Questions

Everything you need to know about deploying inference endpoints on TIR.

Inference & Endpoints

Start Building Today

Your model could be an API by lunch

OpenAI-style routes, Indian GPUs, INR billing.