llmcloud.ai
Apps · Coding agents

Kilo Code

Open-source agent merging Cline and Roo workflows.

Tokens / 7d
141B
Routing directive
auto:coding
On our metal
26%
Licence
Open source

Kilo Code targets cost-sensitive teams and leans hard on open-weight coders. A meaningful share of its traffic lands on llmcloud-hosted DeepSeek and Qwen deployments where the token price is the raw compute price.

How it connects

  • OpenAI-compatible endpoint with per-workspace keys.
  • Cheap model for autocomplete, frontier model for the apply step.
  • Usage attribution via the X-Title header so spend shows per repo.

Traffic pattern

Cost-first. Most turns land on open-weight models; only the final apply escalates to a frontier coder.

claude-sonnet-4.5deepseek-v4grok-code-fast-1

Point it at the gateway

curl https://api.llmcloud.ai/v1/chat/completions \
  -H "Authorization: Bearer $LLMCLOUD_KEY" \
  -H "X-Title: kilo-code/acme-web" \
  -d '{"model":"auto:coding","messages":[...]}'