Coding plans
Monthly plans for coding agents
Subscribe, mint a plan key, paste it into your agent, and code against DeepSeek V4.1 Flash, GLM-5.3 Flash and Qwen3.8 Flash-Next. One flat monthly price: pick a ceiling, not a meter.
Go
Pro
PopularMax
Ultra
Usage is metered as the rupee value of what you use at pay-as-you-go rates, across three rolling windows: 5-hour, weekly and monthly. Plan keys cover LLM chat models only; anything else bills to your wallet at pay-as-you-go rates, on the same GST invoice.USD rates are converted at ₹94.96 = $1 (ECB reference rate, 2026-10-16) and plan prices are rounded. All charges are billed in INR.
One base URL. Your agent, unchanged.
Works with your agent
Point Claude Code, opencode or any OpenAI-compatible coding agent at the same base URL. Your key, your tools, unchanged.
Caching stretches the ceiling
Automatic prompt caching counts cached input at the cached rate, so agent sessions stretch much further than the sticker numbers suggest.
Same sovereignty guarantees
Prices in INR, served entirely from India, zero retention: your diffs and prompts are metered, never stored or trained on.
# point any OpenAI-compatible agent at IndieRouter
base_url = "https://api.indierouter.ai/v1"
# plan models
indierouter/deepseek-v4-flash
indierouter/glm-5.3-flash
indierouter/qwen3.8-flash-nextQuestions
Answers before you commit
How do the rolling windows work?
Each plan includes usage value in three rolling windows: a 5-hour window, a weekly window and a monthly window, all metered as the rupee value of what you use at pay-as-you-go rates. Once a window's included value is used up, the plan key is out of headroom until that window rolls forward; your standard pay-as-you-go key keeps working throughout.
Does prompt caching stretch my quota?
Yes. Prompt caching is automatic, and cached input counts at the cached rate rather than the full input rate. Coding agents resend long context on nearly every turn, so in practice sessions run much further than the window numbers alone suggest.
Do plans auto-renew?
Not yet. A purchase adds 30 days of plan time, and renewing extends it. Nothing charges on its own, and there is no mandate to cancel.
What do plan keys cover?
Chat completions on LLM chat models, nothing else. Anything outside that bills to your prepaid wallet at pay-as-you-go rates. Your standard API keys are unaffected and stay pay as you go; see
Where does my code go?
To GPUs in Indian data centres, exclusively, and nowhere after that: request and response bodies are discarded once the response is served. We log metadata only, for billing and reliability.