llmcloud.ai
Apps · Coding agents

Cline

Autonomous coding agent inside VS Code.

Tokens / 7d
412B
Routing directive
auto:coding
On our metal
18%
Licence
Open source

Cline plans a change, edits files, runs the test suite and iterates until it is green. It is the single largest source of coding tokens on the gateway because each task is a multi-turn loop with the full diff context resent on every step.

How it connects

  • Users paste an llmcloud key into Cline's OpenAI-compatible provider slot.
  • Base URL https://api.llmcloud.ai/v1 with model auto:coding or a pinned model id.
  • Prompt caching cuts the repeated system + repo-map prefix on every loop step.

Traffic pattern

High input-to-output ratio (roughly 20:1). Cache hit rate is the dominant cost lever — a warm prefix cuts a typical task from dollars to cents.

claude-sonnet-4.5gpt-5-codexdeepseek-coder-v3

Point it at the gateway

# Cline → Settings → API Provider: OpenAI Compatible
Base URL: https://api.llmcloud.ai/v1
API Key:  llmc_...
Model ID: auto:coding