Search Traffic is a module inside Botlord Platform. Install and activate the platform first, then configure the engines available to your plan.

Advanced options

SERP exploration, AI features, stealth score, adaptive rate, and scaling to VPS.

3 min read

SERP exploration

SERP Exploration scrolls and hovers over non-target results before clicking the target. It simulates a real user reading the page before choosing a result.

Optionally enable competitor visits — the bot opens a competing result, browses briefly, then returns to search results and clicks your target. This makes the session look like comparison shopping.

Enable exploration when you want higher-quality signals. Disable it for faster, higher-volume runs where session quality matters less than throughput.

AI features

Search Traffic has two distinct AI capabilities. They are different features, enabled differently, and suit different use cases.

Focused Search Traffic provider and model settings: Local LLM is selected and no installed model is chosen. The hint explains the local OpenAI-compatible connection. No connection-test result or ready campaign AI is shown. Select image to enlarge and browse.

AI Assisted mode

AI Assisted mode adds an LLM layer on top of a standard search task. The bot follows the normal route, but before clicking it sends the SERP composition — ad density, AI Overview presence, Maps packs, competing results — to the AI, which picks the best result and timing.

No AI Agent task prompt is required for Assisted mode. Configure a ready provider and inspect the campaign’s AI selection before use. Calls, retries, latency and costs vary by configuration; a provider test or enabled setting does not guarantee improved result selection.

AI Agent tasks

AI Agent tasks are separate from Assisted result selection. Select the task in the campaign builder and write a prompt for supported browsing actions, with a ready provider and the relevant campaign AI settings.

For example, request research of SEO tools and inspection of relevant results. The prompt is an instruction to attempt, not evidence of a unique or completed session; inspect actual actions in Logs.

Model usage and duration depend on the task, provider and retries. Do not infer successful clicks or avoidance of detection. Fill wizard is a separate drafting helper, not an executed AI Agent task.

Providers

OpenAI GPT-4o is the default. Anthropic Claude handles complex multi-step prompts well. OpenRouter routes to many models from one key — useful for cost control. DeepSeek and Gemini are lower-cost options.

Local servers (Ollama, LM Studio) work offline but smaller models often fail multi-step navigation instructions. Use 7B+ and test manually before leaving unattended.

Cloud campaign sync

Use the campaign sync controls in your installed Botlord Platform app to share saved campaigns with your account.

Check the sync status before switching computers. If a sync fails, retry when connected and confirm the expected campaign is present before running it.

Logs and troubleshooting

The Logs tab shows each session step: engine navigation, SERP render, scroll, click, dwell, exit.

Filter by worker ID when running multiple concurrent workers. Screenshots on error can be enabled to capture exactly where a session failed.

Search Traffic startup logs and diagnostic filters; no live session results are shown. Select image to enlarge and browse.
Screenshots is empty in this demo. No live campaign was launched. Select image to enlarge and browse.
Today filters URL history. This empty demo does not demonstrate successful traffic. Select image to enlarge and browse.

Updates

Use Check for Updates in Help & Docs. The platform auto-updates on the stable channel. Beta builds are a separate channel — only install if you are actively testing pre-release features.

After any update, run a short manual session before leaving anything unattended on a VPS.

Help & Docs contains the installed version and update controls. This capture is from build 1.0.592. Select image to enlarge and browse.

Scaling to VPS

For 24/7 or high-concurrency runs, use a Windows VPS with at least 8 GB RAM. The Electron shell, browser profiles, and the antidetect agent together consume significant memory.

Keep an active RDP session open on Windows Server — non-headless browser profiles need a desktop session.

Start with 2–4 workers on a new VPS. Scale up after confirming stable sessions, funded solver balance, and proxy rotation working correctly.

← All topics for Search Traffic