Skip to content

Start free with 1,000 credits - no credit card requiredSee plans

VelATL
Log inStart Free

VelATL + Claude Code

Claude Code, on your VelATL balance.

Point ANTHROPIC_BASE_URL at VelATL and keep your workflow: Messages shape, tool_use blocks, streaming - billed from the same credits as the web app, with per-key spend caps.

Start free with 1,000 creditsRead the API docs

Setup

Claude Code - environment

export ANTHROPIC_BASE_URL="https://velatl.ai/api/v1"
export ANTHROPIC_API_KEY="velatl_..."  # Settings -> API keys, copy once
# model: balanced | fast (see /api/v1/models for your plan's list)

Sanity check - Messages shape

curl https://velatl.ai/api/v1/messages \
  -H "Content-Type: application/json" \
  -H "x-api-key: velatl_..." \
  -H "anthropic-version: 2023-06-01" \
  -d '{"model":"balanced","max_tokens":8192,"messages":[{"role":"user","content":"Hi"}]}'
# Anthropic tools work too: tools[] + tool_choice {"type":"auto"};
# replies stream as tool_use blocks with stop_reason "tool_use".

What works - and what is rejected

  • Tools, streaming and sampling params supported; agentic loops bill per token like chat.
  • Not served: image/audio inputs, thinking blocks, document blocks - text and tools only.
  • stop_sequences and response_format are rejected on this protocol; provider sampling knobs run defaults.
  • Anything unsupported fails with a clear 400 before anything is billed; failed calls refund automatically.
  • Out of credits → 402 insufficient_credits - top up in billing.

Why run it here

Spend caps for agents

Agentic coding loops can burn context fast. Per-key caps plus per-request ledger entries keep every run accountable.

INR + workspace included

UPI and cards in rupees, credits that never expire, and a full web workspace on the same money - not just raw tokens.

Prefer the OpenAI shape instead? See VelATL + OpenCode and the OpenAI-compatible API guide. Full protocol reference: developer docs.

Start free with 1,000 creditsCreate an API key