Files
Moltbot/src/commands/vllm-setup.ts
2026-02-14 15:39:46 +00:00

79 lines
2.4 KiB
TypeScript

import type { OpenClawConfig } from "../config/config.js";
import type { WizardPrompter } from "../wizard/prompts.js";
import { upsertAuthProfileWithLock } from "../agents/auth-profiles.js";
export const VLLM_DEFAULT_BASE_URL = "http://127.0.0.1:8000/v1";
export const VLLM_DEFAULT_CONTEXT_WINDOW = 128000;
export const VLLM_DEFAULT_MAX_TOKENS = 8192;
export const VLLM_DEFAULT_COST = {
input: 0,
output: 0,
cacheRead: 0,
cacheWrite: 0,
};
export async function promptAndConfigureVllm(params: {
cfg: OpenClawConfig;
prompter: WizardPrompter;
agentDir?: string;
}): Promise<{ config: OpenClawConfig; modelId: string; modelRef: string }> {
const baseUrlRaw = await params.prompter.text({
message: "vLLM base URL",
initialValue: VLLM_DEFAULT_BASE_URL,
placeholder: VLLM_DEFAULT_BASE_URL,
validate: (value) => (value?.trim() ? undefined : "Required"),
});
const apiKeyRaw = await params.prompter.text({
message: "vLLM API key",
placeholder: "sk-... (or any non-empty string)",
validate: (value) => (value?.trim() ? undefined : "Required"),
});
const modelIdRaw = await params.prompter.text({
message: "vLLM model",
placeholder: "meta-llama/Meta-Llama-3-8B-Instruct",
validate: (value) => (value?.trim() ? undefined : "Required"),
});
const baseUrl = String(baseUrlRaw ?? "")
.trim()
.replace(/\/+$/, "");
const apiKey = String(apiKeyRaw ?? "").trim();
const modelId = String(modelIdRaw ?? "").trim();
const modelRef = `vllm/${modelId}`;
await upsertAuthProfileWithLock({
profileId: "vllm:default",
credential: { type: "api_key", provider: "vllm", key: apiKey },
agentDir: params.agentDir,
});
const nextConfig: OpenClawConfig = {
...params.cfg,
models: {
...params.cfg.models,
mode: params.cfg.models?.mode ?? "merge",
providers: {
...params.cfg.models?.providers,
vllm: {
baseUrl,
api: "openai-completions",
apiKey: "VLLM_API_KEY",
models: [
{
id: modelId,
name: modelId,
reasoning: false,
input: ["text"],
cost: VLLM_DEFAULT_COST,
contextWindow: VLLM_DEFAULT_CONTEXT_WINDOW,
maxTokens: VLLM_DEFAULT_MAX_TOKENS,
},
],
},
},
},
};
return { config: nextConfig, modelId, modelRef };
}