> For the complete documentation index, see [llms.txt](https://seekspeed.gitbook.io/seekspeed-docs/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://seekspeed.gitbook.io/seekspeed-docs/getting-started/quickstart.md).

# Quickstart

A five-minute path from cold start to your first statistically significant benchmark result.

## 1. Add a connector

Navigate to `/app/connectors` and pick a preset. Each preset injects the right base URL, model field, and provider-specific headers:

```ts
// src/lib/connector-presets.ts (excerpt)
export const PRESETS = {
  openai:    { baseUrl: "https://api.openai.com/v1", model: "gpt-4o-mini" },
  groq:      { baseUrl: "https://api.groq.com/openai/v1", model: "llama-3.1-70b-versatile" },
  together:  { baseUrl: "https://api.together.xyz/v1", model: "meta-llama/Llama-3-70b-chat-hf" },
  vllm:      { baseUrl: "http://localhost:8000/v1", model: "deepseek-ai/DeepSeek-V3" },
  tgi:       { baseUrl: "http://localhost:8080/v1", model: "deepseek-ai/DeepSeek-V3" },
  llamacpp:  { baseUrl: "http://localhost:8080/v1", model: "deepseek-coder" },
};
```

Click **Test** before save. The server fn fires a 1-token completion and reports status + latency:

```ts
// src/lib/connector-test.functions.ts
export const testConnector = createServerFn({ method: "POST" })
  .inputValidator((d) => z.object({ baseUrl: z.string().url(), apiKey: z.string().optional(), model: z.string() }).parse(d))
  .handler(async ({ data }) => {
    const t0 = Date.now();
    const res = await fetch(`${data.baseUrl}/chat/completions`, {
      method: "POST",
      headers: { "Content-Type": "application/json", ...(data.apiKey && { Authorization: `Bearer ${data.apiKey}` }) },
      body: JSON.stringify({ model: data.model, messages: [{ role: "user", content: "ping" }], max_tokens: 1 }),
    });
    return { ok: res.ok, status: res.status, ms: Date.now() - t0 };
  });
```

## 2. Create an agent profile

```
System: You are a senior engineer. Be terse.
Connector: groq-llama-70b
Tools: [search, fetch_url]
Memory: ephemeral
Token budget: 2048
```

## 3. Run a benchmark

`/app/benchmarks → New benchmark`. Paste a CSV of prompts or use the built-in set. Configure:

* **Iterations**: 5 (per prompt)
* **Concurrency**: 1, 4, 16 (sweep)
* **Stream**: on (for real TTFT)

Hit **Run**. Recharts draws three lines as results stream in: TTFT, total latency, throughput.

## 4. Read the recommendation

The optimization engine inspects every completed run:

```ts
// src/lib/recommendations.ts (excerpt)
if (avgPromptTokens > 1500 && p95LatencyMs > 4000) {
  push({
    title: "Slim system prompt",
    rationale: `Avg prompt is ${avgPromptTokens} tokens; p95 latency is ${p95LatencyMs}ms.`,
    expectedImpact: "20-35% TTFT reduction",
    difficulty: "low",
    confidence: 0.82,
    apply: variantWithSlimmedPrompt,
  });
}
```

## 5. Apply, re-run, accept

Click **Apply** on a recommendation. SeekSpeed generates an agent variant, re-runs the benchmark, and surfaces the delta with a Welch's *t*-test. If `p < 0.05` and the mean improved, **Accept** promotes the variant. Otherwise **Reject** and the variant is archived.

Continue to [Core Concepts →](/seekspeed-docs/getting-started/core-concepts.md)


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter:

```
GET https://seekspeed.gitbook.io/seekspeed-docs/getting-started/quickstart.md?ask=<question>&goal=<endgoal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.