Connect OpenCode to Token Factory

​

Prerequisites

• A Token Factory API key. See Authentication.
• OpenCode 1.19.0 or later.
• A model from the Token Factory catalog.

​

Context window and compaction

• limit.context is the model’s context_window.
• limit.output is the model’s max_output_tokens. OpenCode sends max_tokens = the smaller of this and its built-in 32,000-token cap.
• limit.input sets where OpenCode compacts: it summarizes the conversation once a request reaches limit.input minus 20,000 tokens (its default compaction.reserved while the output cap is at least 20,000). Without limit.input, OpenCode compacts only at limit.context − max_tokens, with no headroom.

export CORVEX_API_KEY="sk-corvex-..."

{
  "$schema": "https://opencode.ai/config.json",
  "provider": {
    "corvex": {
      "npm": "@ai-sdk/openai-compatible",
      "name": "Corvex Token Factory",
      "options": {
        "baseURL": "https://api.tokenfactory.corvex.cloud/v1",
        "apiKey": "{env:CORVEX_API_KEY}"
      },
      "models": {
        "zai-org/GLM-5.3": {
          "name": "GLM 5.3",
          "reasoning": true,
          "tool_call": true,
          "cost": { "input": 1.4, "output": 4.4 },
          "limit": { "context": 393216, "input": 334100, "output": 393216 }
        },
        "deepseek-ai/DeepSeek-V4-Flash-0731": {
          "name": "DeepSeek V4 Flash 0731",
          "reasoning": true,
          "tool_call": true,
          "cost": { "input": 0.14, "output": 0.28 },
          "limit": { "context": 1048576, "input": 903979, "output": 1048576 }
        }
      }
    }
  },
  "model": "corvex/zai-org/GLM-5.3"
}

opencode run -m corvex/deepseek-ai/DeepSeek-V4-Flash-0731 "Summarize this repository."

​

Compatibility notes

• Output budget. Reasoning tokens count toward the 32,000-token step budget. If a response or tool call is truncated, retry with a shorter prompt or split the task. The OPENCODE_EXPERIMENTAL_OUTPUT_TOKEN_MAX environment variable changes the cap; recompute the compaction window if you change it.
• Context length. If a session still exceeds the model’s window, the gateway returns context_length_exceeded. Run /compact or start a new session. See Errors.
• Model IDs. Use the exact id from GET /v1/models, prefixed with the provider ID (corvex/) when you select it on the command line.

​

Related documentation

• Connect Kilo Code — the same configuration format in Kilo Code.
• Connect pi and Connect Codex — other terminal coding agents.
• Errors — authentication, rate-limit, and validation responses.