Replymer monitors social media platforms (X and Reddit) for mentions that match your goals, scores them for relevance, and publishes AI-generated replies through your managed browser accounts — all automatically.

Key concepts

• Task — the primary unit of work. Defines what to monitor, where, how often, and what to do with findings.
• Mention — a social post that matched your task keywords, scored for relevance.
• Reply — an AI-generated response draft, awaiting your approval or published automatically.
• Managed Account — a connected social account (X or Reddit) used to post replies.
• Workspace — your account namespace. API keys are workspace-scoped.

Platforms

• X — search API + scraping via managed browser profiles
• Reddit — Reddit search API, subreddit monitoring

Get your first task running in under 5 minutes.

1. Create an account

Sign up at replymer.ai. You receive 3 free trial replies to get started — no credit card required.

2. Add a managed account

Go to Settings → Accounts and connect an X or Reddit account. Replymer uses browser automation to post replies on your behalf — your account must be active and in good standing.

3. Create your first task

Click New Task in the sidebar (or ask the AI Agent). Choose a goal, add keywords that describe your product or the conversations you want to find, and pick the platforms to monitor.

{
  "name": "Lead Gen — SaaS Founders",
  "goal": "lead_generation",
  "platforms": ["twitter", "reddit"],
  "keywords": [
    "project management tool",
    "trello alternative",
    "asana vs notion"
  ],
  "scheduleFrequency": "hourly",
  "relevanceThreshold": 65,
  "replyTone": "helpful"
}

4. Review mentions

Once the first scan runs (within a minute of creating the task), relevant mentions appear in the Mentions feed. Each mention has a relevance score (0–100) and an AI explanation.

5. Approve and publish replies

AI-generated reply drafts appear in the Replies queue. Review, edit if needed, and approve — or enable Auto-Reply on the task to publish automatically.

The public API uses Bearer token authentication. Generate an API key from the dashboard and pass it in every request header.

Authorization: Bearer rk_your_api_key_here

Creating an API key

1. Open Settings → API Keys in the dashboard
2. Click Create Key and give it a name
3. Copy the key immediately — it is shown only once
4. Store it securely (environment variable, secret manager)

Key format

All keys start with rk_ followed by a random string. In the dashboard you only see the prefix (e.g. rk_a1b2c3••••••••) — the full key is stored as a SHA-256 hash.

Revoking a key

Go to Settings → API Keys, click the trash icon next to the key. All subsequent requests with that key return 401 Unauthorizedimmediately.

Authentication errors

{
  "error": {
    "code": "UNAUTHORIZED",
    "message": "Missing or invalid API key"
  }
}

A task defines what to monitor, where, and why. The goal field shapes how the AI scores relevance — the same mention might score 90 for a lead-gen task but 20 for a brand-monitoring task.

Goals

Keywords

Keywords are the raw search terms sent to platform search APIs. Use specific, meaningful phrases — not single words. Examples:

• ✅ “project management tool for remote teams”
• ✅ “trello alternative with time tracking”
• ❌ “project” — too generic, returns too much noise

Relevance threshold

Only mentions with a relevance score ≥ threshold appear in the feed. Default is 60. Lower = more mentions (higher noise), higher = fewer but more targeted mentions.

Schedule

You can optionally restrict scanning to active hours (e.g. 9am–6pm in your timezone) to avoid replying to old posts during off-hours.

Auto-Reply

When enabled, Replymer automatically publishes replies for all relevant mentions above the threshold without waiting for manual approval. Set autoReplyAccountId to specify which managed account to use.

When a scan runs, it fetches posts matching your keywords from the target platforms, then the AI scores each one for relevance to your task goal. Mentions above your threshold appear in the feed.

Status lifecycle

Relevance score

The score (0–100) is computed by the AI with the task goal as context. A score of 90+ means the post is highly relevant — the author is likely a perfect prospect or direct brand mention. The AI also provides a short relevanceReason explaining the score.

Sentiment

Each mention is classified as positive, negative, neutral, or question. Questions and negatives are often the highest-value targets for lead-gen and support tasks.

For each relevant mention, Replymer generates a reply draft using the task's tone and instructions. Drafts go through quality scoring before appearing in the approval queue.

Quality scoring

• Quality score (0–100) — measures how helpful, natural, and on-topic the reply is
• Spam risk score (0–100) — measures how promotional or spammy the reply looks; lower is better
• Naturalness score (0–100) — measures how human and authentic the reply sounds; higher is better

Replies with spam risk > 60 or naturalness < 40 are automatically discarded. Auto-approved replies require quality ≥ 70, spam ≤ 30, and naturalness ≥ 60.

Reply status lifecycle

Tone options

• helpful — friendly, informative, problem-solving
• casual — conversational, approachable
• professional — formal, business-appropriate
• technical — precise, detailed, data-driven
• empathetic — warm, understanding, supportive
• witty — sharp, slightly humorous
• custom — provide your own instructions via replyInstructions

Scans are the execution units — each scan runs the search for one platform and one task. A task with two platforms creates two scan jobs per scheduled run.

Scan pipeline

1. BullMQ job is created for each platform at schedule time
2. Platform search API is queried with task keywords
3. Raw results are deduped against previously seen posts
4. AI scores each new post (batched for efficiency)
5. Relevant mentions are saved and appear in the feed
6. Reply generation is queued for all relevant mentions (if not manual-only)

Cost

Scans and AI scoring are free — included with every plan. You only pay for published replies that exceed your monthly plan limit.

Manual scans

Trigger an immediate scan via the dashboard or the API. Useful after updating keywords or for testing. Manual scans run in addition to the schedule.

Replymer publishes replies via browser automation (AdsPower profiles). Each managed account represents one social media account (X or Reddit) that is linked to your workspace.

Adding an account

1. Go to Settings → Accounts
2. Click Add Account
3. Enter the platform, username, and AdsPower profile ID
4. Replymer will verify the account is accessible and set status to active

Account health

• active — ready to post
• rate_limited — platform is throttling; Replymer backs off automatically
• suspended — account banned; manual intervention required
• error — browser profile issue; check AdsPower

Replymer uses a simple reply-based pricing model. Scanning and AI generation are free with every plan — you only pay for published replies.

How it works

• Each plan includes a monthly reply allowance (e.g. Growth = 200 replies/month)
• Scans, AI scoring, and reply generation are unlimited and free
• Only successfully published replies count against your limit
• Overage replies are charged from your overage credit balance at your plan's rate

Plans

Overage credits

If you exceed your monthly reply limit, additional replies are charged from your overage credit balance. Top up any time from the Billing page — available in $10, $25, $50, and $100 increments. Credits never expire.

Free trial

Every new workspace receives 3 free trial replies automatically — no credit card required. Use them to test the full pipeline before subscribing.

Response format

{ "data": { ... } }

{
  "items": [ ... ],
  "hasMore": true,
  "nextCursor": "2024-01-15T12:00:00.000Z"
}

{
  "error": {
    "code": "NOT_FOUND",
    "message": "Task not found"
  }
}

HTTP status codes

Pagination

List endpoints use cursor-based pagination. Pass nextCursor from the previous response as the cursor query param in the next request.

GET /api/v1/mentions?limit=50&cursor=2024-01-15T12:00:00.000Z

Returns workspace info, reply usage, overage balance, and task counts.

curl https://replymer.ai/api/v1/workspace \
  -H "Authorization: Bearer rk_your_key"

{
  "data": {
    "id": "clxyz...",
    "name": "My Workspace",
    "slug": "my-workspace",
    "balanceUsd": 10.00,
    "totalSpentUsd": 24.00,
    "subscriptionPlan": "growth",
    "repliesUsedThisPeriod": 42,
    "monthlyReplyLimit": 200,
    "freeTrialRepliesLeft": 0,
    "overageRateUsd": 0.40,
    "defaultReplyTone": "helpful",
    "defaultReplyLanguage": "en",
    "tasks": { "active": 3, "paused": 1 },
    "last24hScans": 48
  }
}

List all tasks. Returns non-removed tasks by default. Filter with ?status=paused.

curl "https://replymer.ai/api/v1/tasks?status=active" \
  -H "Authorization: Bearer rk_your_key"

Create a new task.

curl -X POST https://replymer.ai/api/v1/tasks \
  -H "Authorization: Bearer rk_your_key" \
  -H "Content-Type: application/json" \
  -d '{
    "name": "Lead Gen — SaaS Founders",
    "goal": "lead_generation",
    "platforms": ["twitter", "reddit"],
    "keywords": ["project management tool", "trello alternative"],
    "scheduleFrequency": "hourly",
    "relevanceThreshold": 65
  }'

Get a single task including mention and scan counts.

Update task settings. Send only the fields you want to change.

curl -X PATCH https://replymer.ai/api/v1/tasks/task_abc123 \
  -H "Authorization: Bearer rk_your_key" \
  -H "Content-Type: application/json" \
  -d '{ "status": "paused" }'

Soft-delete a task (sets status: “removed”). Data is preserved and queryable with ?status=removed.

List mentions with filtering and cursor-based pagination.

curl "https://replymer.ai/api/v1/mentions?taskId=task_abc&minScore=80&platform=reddit" \
  -H "Authorization: Bearer rk_your_key"

Get a single mention with its replies.

Update a mention's status. Valid values: relevant, irrelevant, skipped, queued.

curl -X PATCH https://replymer.ai/api/v1/mentions/mention_xyz \
  -H "Authorization: Bearer rk_your_key" \
  -H "Content-Type: application/json" \
  -d '{ "status": "skipped" }'

List reply drafts with filtering and pagination.

curl "https://replymer.ai/api/v1/replies?status=draft" \
  -H "Authorization: Bearer rk_your_key"

Get a single reply with its mention context.

Approve, reject, or edit a reply draft.

curl -X PATCH https://replymer.ai/api/v1/replies/reply_123 \
  -H "Authorization: Bearer rk_your_key" \
  -H "Content-Type: application/json" \
  -d '{ "action": "approve" }'

# First edit the content
curl -X PATCH https://replymer.ai/api/v1/replies/reply_123 \
  -H "Authorization: Bearer rk_your_key" \
  -H "Content-Type: application/json" \
  -d '{
    "action": "edit",
    "content": "Great question! We built this exact feature in v2..."
  }'

# Then approve
curl -X PATCH https://replymer.ai/api/v1/replies/reply_123 \
  -H "Authorization: Bearer rk_your_key" \
  -H "Content-Type: application/json" \
  -d '{ "action": "approve" }'

Trigger a manual scan for a task. Returns 202 Accepted immediately — scanning happens asynchronously.

curl -X POST https://replymer.ai/api/v1/scan \
  -H "Authorization: Bearer rk_your_key" \
  -H "Content-Type: application/json" \
  -d '{ "taskId": "task_abc123" }'

{
  "data": {
    "taskId": "task_abc123",
    "jobs": [
      { "platform": "twitter", "jobId": "scan-task_abc123-twitter-1705312800000" },
      { "platform": "reddit",  "jobId": "scan-task_abc123-reddit-1705312800000" }
    ]
  }
}

List recent scan executions.

List all active API keys for the workspace. Returns only the prefix — the full key is never stored or returned after creation.

Create a new API key. The full key is returned exactly once.

{
  "id": "key_abc",
  "name": "My automation",
  "keyPrefix": "rk_a1b2c3",
  "key": "rk_a1b2c3d4e5f6...",
  "createdAt": "2024-01-15T09:00:00.000Z"
}

Revoke an API key immediately. All subsequent requests with that key return 401.

// Trigger scan
await fetch("https://replymer.ai/api/v1/scan", {
  method: "POST",
  headers: {
    "Authorization": "Bearer rk_your_key",
    "Content-Type": "application/json",
  },
  body: JSON.stringify({ taskId: "task_abc123" }),
});

// Wait ~30s for scan to complete, then fetch results
await new Promise(r => setTimeout(r, 30_000));

const { items } = await fetch(
  "https://replymer.ai/api/v1/mentions?taskId=task_abc123&status=relevant&minScore=70",
  { headers: { "Authorization": "Bearer rk_your_key" } }
).then(r => r.json());

let cursor = null;

do {
  const url = new URL("https://replymer.ai/api/v1/replies");
  url.searchParams.set("taskId", "task_abc123");
  url.searchParams.set("status", "draft");
  if (cursor) url.searchParams.set("cursor", cursor);

  const { items, hasMore, nextCursor } = await fetch(url, {
    headers: { "Authorization": "Bearer rk_your_key" },
  }).then(r => r.json());

  for (const reply of items) {
    await fetch(`https://replymer.ai/api/v1/replies/${reply.id}`, {
      method: "PATCH",
      headers: {
        "Authorization": "Bearer rk_your_key",
        "Content-Type": "application/json",
      },
      body: JSON.stringify({ action: "approve" }),
    });
  }

  cursor = hasMore ? nextCursor : null;
} while (cursor);