Generate, review, and approve Reddit posts via API or Zapier — no dashboard needed.
API Endpoints
Generating Reddit content at scale means either hiring a writer who understands Reddit's culture (expensive and rare), or manually writing everything yourself (not scalable). The AI alternatives — just prompting ChatGPT with "write a Reddit post" — produce content that Reddit readers immediately identify as inauthentic, resulting in downvotes and bans.
Even when AI-generated content is high quality, the review workflow is broken. Posts get generated, sit in a folder, never get reviewed because nobody built a process, and eventually the whole campaign stalls. The bottleneck isn't generation — it's the approval pipeline.
The solution is a structured API-driven pipeline: AI generates posts optimised for specific subreddits, a queue surfaces them for review, humans approve or reject with a single API call, and approved posts go straight into the scheduler. The AI handles volume; humans handle quality control; the pipeline handles everything else.
The Reoogle AI generates Reddit posts tailored to each target community — using the subreddit's culture, posting conventions, and active topics as context. The /ai-posts endpoints give you programmatic control over the entire approval workflow: pull pending posts, inspect the AI's analysis, edit any field before approving, and approve or reject with a single API call. Wire this into a Zapier automation, a Slack approval bot, or your own dashboard, and you have a content factory that scales without proportionally scaling headcount.
Configure your content profile
Ensure your business profile has business_name, website_url, and target_category set in the dashboard. The AI uses these to generate posts that naturally reference your product and target the right subreddit categories. This is a one-time setup step.
Poll for newly generated posts
Call GET /ai-posts?status=pending&per_page=50 on a schedule — every 15–30 minutes during active hours is typical. Each post includes the AI-generated title, body, target subreddit, and an optimization_score (0–100) reflecting how well the post fits the community.
Use the optimization score to triage
Posts with optimization_score ≥ 70 are strong candidates for approval. Posts below 50 likely need heavy editing or rejection. Build your review workflow around this score: auto-queue high scores for fast review, route low scores to a detailed editing step.
Read the AI analysis before approving
Each post includes an ai_analysis field: 2–3 sentences explaining why the post fits the target subreddit, what posting strategy it uses, and what makes it likely to perform. Read this before approving — it tells you exactly what the AI was optimising for and highlights any risks.
Edit before approving if needed
Use PUT /ai-posts/{id} to update the title, body, or switch the subreddit. All fields are optional — send only what you're changing. Common edits: making the title more specific, adding a personal detail to the body, or swapping to a more niche subreddit for better targeting.
Approve or reject
POST /ai-posts/{id}/approve queues the post for immediate Reddit submission via the connected account. POST /ai-posts/{id}/reject discards it. Rejected posts won't be re-generated — if a post has potential but needs more work, edit it rather than rejecting.
Feed performance data back into the system
After 48 hours, posts have real Reddit performance data: score, upvote_ratio, comment_count. Export these via GET /ai-posts?status=posted and identify which generated posts outperformed. Note the common patterns — subreddit, title length, body structure, tone — and include them in your product's content brief to steer the AI toward better outputs.
Fetch and auto-approve high-quality posts
const API = 0 ;
const HEADERS = { Authorization: 1 };
interface AiPost {
id: string; title: string; body: string;
subreddit_name: string; optimization_score: number;
}
async function processPendingPosts(minScore = 70) {
const res = await fetch( 2 , { headers: HEADERS });
const posts = (await res.json()).data as AiPost[];
for (const post of posts) {
if (post.optimization_score >= minScore) {
3
await fetch( 4 , { method: 5 , headers: HEADERS });
console.log( 6 );
} else {
7
await fetch( 8 , { method: 9 , headers: HEADERS });
console.log( 10 );
}
}
}Edit a post before approving
# Update title and body, then approve
curl -X PUT https://reoogle.com/api/v1/ai-posts/POST_UUID \
-H "Authorization: Bearer rog_live_YOUR_KEY" \
-H "Content-Type: application/json" \
-d '{
"title": "Revised: How we grew 0 → 1k users without paid ads",
"body": "Updated body with better hook..."
}'
# Then approve it
curl -X POST https://reoogle.com/api/v1/ai-posts/POST_UUID/approve \
-H "Authorization: Bearer rog_live_YOUR_KEY"Auto-approving posts without reading the AI analysis
The optimization_score is a signal, not a guarantee. A score of 80 on the wrong subreddit, or with a title that slightly misreads the community tone, will still underperform. Always read ai_analysis before approving.
Approving faster than you can review quality
The temptation is to set a score threshold and auto-approve everything above it. This creates a fire-and-forget workflow that eventually produces a bad post, damages your account's standing in a subreddit, and is hard to undo. Keep a human in the loop.
Ignoring rejected post patterns
Every rejected post is a data point. If you're rejecting most AI outputs for the same reason (too salesy, wrong tone for that subreddit, too long), update your content brief in the dashboard — don't just reject in silence.
4 min
Review time per post
Median time for an experienced reviewer to read, lightly edit, and approve/reject an AI-generated post.
5×
Content volume increase
Typical increase in weekly post volume for a content team that shifts from fully manual to AI + human review.
74%
Approval rate for high-score posts
Percentage of posts with optimization_score ≥ 70 that are approved without edits.
How does the AI know which subreddit to target for each post?
The AI combines your product category, website URL, and recent high-performing subreddits from your history to select a target community. You can override this by specifying subreddit_name when the post is in pending state via PUT /ai-posts/{id}.
Can I build a Slack bot that handles approvals without a custom dashboard?
Yes — this is one of the most popular integrations. Poll /ai-posts?status=pending via Zapier, send each post as a Slack message with Approve/Reject buttons, and have the button actions call the approve or reject endpoints. No custom UI needed.
What's the difference between "pending" and "pending_review"?
"pending" means the post was generated by the AI and is waiting for approval. "pending_review" means the post was submitted to Reddit and is in a subreddit's queue waiting for mod approval. Both require different handling — pending needs your review; pending_review just needs patience.
Can I set up automatic approval without human review?
Yes — just call POST /ai-posts/{id}/approve for any post with a score above your threshold. We recommend running auto-approval only during a pilot period after you've reviewed 30–50 posts manually and established that the AI's output quality is consistently acceptable for your use case.
Business value
A content manager spending 3 hours/day on Reddit posts can reduce that to 30 minutes of review using this pipeline. The AI handles generation and scheduling; the human just approves. One person can manage 5× the content volume.
Generate an API key from your developer dashboard and make your first request in minutes.