1
Get the key
Copy the delivered key from Dashboard. Do not place the raw key in a public repository or commit it into a configuration file.
Agent Plan / Coding Plan
Coding Plan Access & Integration
Follow the same practical order used by the provider documentation: start with a model call, choose the correct access method, connect your coding tool, and verify the active route.
Current stocked access
Agent Coding Access & Configuration Package — One-Time
Upstream key type: Volcengine Ark Agent Plan Medium. This page explains how to call the route directly and how to connect the supported coding tools.
Quota reference
GLM 5.2 / DeepSeek V4
Quota windows
≈20M / 5H · ≈78M / 1 Week · ≈220M / 1 Month
Default route
ark-code-latest
DeepSeek V4GLM 5.2MiniMax M3Kimi K2.6
1 · Quick start
Start by calling the model
The key is a direct access credential. Put it in the tool or environment variable required by the client, then call the provider route with the default model route below.
2
Choose the protocol
Use the Anthropic-compatible route for Claude Code. Use the /v3 OpenAI-compatible route for Codex, OpenCode, OpenClaw, and Hermes.
3
Verify the route
Keep
ark-code-latest in the client configuration, then switch the enabled model upstream in Volcengine when needed.2 · Model call
Test the OpenAI-compatible connection
This test confirms the key, base URL, protocol, and default route before you add a larger coding-agent workflow.
Chat Completions
https://ark.cn-beijing.volces.com/api/plan/v3/chat/completionsResponses API
https://ark.cn-beijing.volces.com/api/plan/v3/responsesClaude Code
https://ark.cn-beijing.volces.com/api/planexport ARK_API_KEY="YOUR_AGENT_PLAN_KEY"
curl https://ark.cn-beijing.volces.com/api/plan/v3/chat/completions \
-H "Authorization: Bearer $ARK_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "ark-code-latest",
"messages": [{"role": "user", "content": "Reply with a short connection confirmation."}]
}'The visible model names are selected upstream. The client route stays ark-code-latest for this Agent Plan connection.
3 · Connect AI tools
Choose the access method for your tool
Each tool has its own protocol, secret name, and configuration file. Start with the matching card; do not copy the Claude Code route into an OpenAI-compatible client.
Codex
OpenAI Responses API
- Configuration file
- ~/.codex/config.toml
- Secret / key location
- ARK_API_KEY
- Start here
- Set ARK_API_KEY, then set the provider base URL and wire_api = responses.
Claude Code
Anthropic-compatible API
- Configuration file
- ~/.claude/settings.json
- Secret / key location
- ANTHROPIC_AUTH_TOKEN
- Start here
- Set the Anthropic token and base URL, then keep the Agent route as the configured model.
OpenCode
OpenAI Responses API
- Configuration file
- ~/.config/opencode/opencode.json
- Secret / key location
- API key in provider.options
- Start here
- Create a provider using @ai-sdk/openai, the /v3 base URL, and ark-code-latest.
OpenClaw
OpenAI-compatible Chat Completions
- Configuration file
- ~/.openclaw/openclaw.json
- Secret / key location
- apiKey in models.providers
- Start here
- Merge the provider and agent model blocks; do not overwrite existing channels or Skills.
Hermes Agent
Custom OpenAI-compatible endpoint
- Configuration file
- ~/.hermes/config.yaml + ~/.hermes/.env
- Secret / key location
- ARK_API_KEY
- Start here
- Use the custom endpoint, keep the route in config.yaml, and store the key in the environment file.
4 · Verify and switch models
Confirm the connection before starting a full project
01
Run one short request and confirm a normal response.
02
Open Agent Plan management and select an enabled model.
03
Wait 3–5 minutes, restart the tool, and run its status command again.