OpenRouter analyze

Direct OpenRouter API proxy for subreddit analysis

POST /api/openrouter/analyze-subreddit

Proxy endpoint for direct OpenRouter LLM analysis. Used by frontend Web Workers for AI-powered subreddit analysis. Unlike /api/analyze/:subreddit, this endpoint does not fetch Reddit data or perform caching—it simply forwards the request to OpenRouter.

Authentication

Requires valid JWT token. No subscription check is performed at this endpoint level (quota is enforced).

Quota Management

This endpoint enforces the analysis quota:

• Checks monthly usage before processing
• Returns 403 if quota exceeded
• Increments counter on success
• Admin users bypass quota limits

Request Body

data · object · required

Subreddit data object to analyze

Show child attributes

data.name · string · required

Subreddit name

data.title · string

Subreddit display title

data.description · string

Public description

data.rules · array

Array of rule objects (short_name, description)

data.content_categories · array

Array of allowed content types

data.requires_approval · boolean

Whether posts require mod approval

data.karma_required · boolean

Whether karma is required to post

data.account_age_required · boolean

Whether account age restrictions exist

systemPrompt · string

Custom system prompt (defaults to Reddit marketing analyst prompt)

responseFormat · object

OpenRouter response format configuration (defaults to {type: "json_object"})

OpenRouter Configuration

From server.js:1499-1565:

const OPENROUTER_MODEL = 'google/gemini-3-flash-preview';

const result = await callOpenRouterChatCompletions({
  apiKey: process.env.OPENROUTER_API_KEY,
  referer: req.headers.origin,
  title: 'SubPirate - Reddit Marketing Analysis',
  body: {
    model: OPENROUTER_MODEL,
    messages: [
      {
        role: 'system',
        content: systemPrompt || 'You are an expert Reddit marketing analyst...'
      },
      { role: 'user', content: prompt }
    ],
    temperature: 0.6,
    max_tokens: 35000,
    stream: false,
    response_format: responseFormat || { type: 'json_object' }
  }
});

Response

Returns the raw OpenRouter API response JSON.

choices · array

OpenRouter response choices

Show child attributes

message · object

LLM message object

Show child attributes

role · string

Always “assistant”

content · string

Raw JSON string with analysis results

finish_reason · string

Reason for completion (e.g., “stop”)

usage · object

Token usage statistics

Show child attributes

prompt_tokens · number

Input tokens consumed

completion_tokens · number

Output tokens generated

total_tokens · number

Total tokens used

Example Request

cURL

curl -X POST "https://api.example.com/api/openrouter/analyze-subreddit" \
  -H "Authorization: Bearer YOUR_JWT_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "data": {
      "name": "startups",
      "title": "Startups",
      "description": "Discussion about startups and entrepreneurship",
      "rules": [
        {
          "short_name": "No spam",
          "description": "No promotional content"
        }
      ],
      "content_categories": [],
      "requires_approval": false,
      "karma_required": false,
      "account_age_required": false
    },
    "systemPrompt": "You are an expert Reddit marketing analyst. Follow subreddit rules.",
    "responseFormat": { "type": "json_object" }
  }'

Example Response

{
  "id": "gen-abc123",
  "model": "google/gemini-3-flash-preview",
  "choices": [
    {
      "message": {
        "role": "assistant",
        "content": "{\"marketingFriendliness\":{\"score\":45,\"reasons\":[...]}}"
      },
      "finish_reason": "stop"
    }
  ],
  "usage": {
    "prompt_tokens": 1250,
    "completion_tokens": 890,
    "total_tokens": 2140
  }
}

Error Responses

400 · error

Missing required data.name field

{
  "error": "Missing required data for subreddit analysis"
}

403 · error

Quota exceeded

{
  "error": "quota_exceeded",
  "feature": "analysis",
  "used": 10,
  "limit": 10,
  "upgrade_required": true,
  "message": "You've used all 10 analyses this month. Upgrade for more."
}

500 · error

OpenRouter API key not configured or internal error

502 · error

OpenRouter API error

{
  "error": "OpenRouter API error: 429",
  "details": "Rate limit exceeded"
}

Usage in Frontend

This endpoint is used by Web Workers for asynchronous analysis:

// From analysisWorker.ts
const response = await fetch('/api/openrouter/analyze-subreddit', {
  method: 'POST',
  headers: {
    'Content-Type': 'application/json',
    'Authorization': `Bearer ${token}`
  },
  body: JSON.stringify({
    data: subredditData,
    systemPrompt: 'You are an expert Reddit marketing analyst...',
    responseFormat: { type: 'json_object' }
  })
});

const result = await response.json();
const analysis = JSON.parse(result.choices[0].message.content);

Prompt Building

The prompt is constructed by buildAnalysisPrompt() (from api/_lib/analyze.js and server.js:586-630):

function buildAnalysisPrompt(data, options = {}) {
  const { topPosts, modSummary } = options;

  let prompt = `Analyze the following subreddit for marketing potential:

Subreddit: r/${data.name}
Title: ${data.title}
Description: ${data.description}

Rules:
${data.rules.map((r, i) => `${i+1}. ${r.short_name || r.title}: ${r.description}`).join('\n')}

Restrictions:
- Posts require approval: ${data.requires_approval}
- Karma requirement: ${data.karma_required}
- Account age requirement: ${data.account_age_required}
`;

  if (topPosts?.length > 0) {
    prompt += `\n\nTop Posts (for title pattern analysis):\n`;
    topPosts.slice(0, 10).forEach(p => {
      prompt += `- "${p.title}" (${p.score} upvotes, ${p.num_comments} comments)\n`;
    });
  }

  if (modSummary) {
    prompt += `\n\nModerator Activity:\n`;
    prompt += `- Total mods: ${modSummary.totalHumanMods}\n`;
    prompt += `- Very active (< 7 days): ${modSummary.veryActive}\n`;
    prompt += `- Active (< 30 days): ${modSummary.active}\n`;
    prompt += `- Dormant (> 30 days): ${modSummary.dormant}\n`;
  }

  return prompt;
}

Rate Limiting

OpenRouter enforces rate limits:

• Free tier: 20 requests/minute, 200 requests/day
• Paid: Higher limits based on credit balance
• Model-specific: Gemini 3 Flash Preview has generous limits

If rate limited, retry with exponential backoff.

Next Steps

Full Analysis Endpoint

Complete analysis with Reddit data fetching and caching

OpenRouter Setup

Configure your OpenRouter API key