OpenRouter AI Engine
Applivo's internal multi-model reasoning gateway routing candidate profiling across state-of-the-art LLMs.

Overview
OpenRouter is the leading unified gateway for cutting-edge artificial intelligence models, routing inference across Anthropic Claude, OpenAI GPT-4o, and Meta Llama architectures. Applivo uses OpenRouter as its central AI reasoning backbone for job match scoring, cover letter drafting, and candidate question formulation.
By decoupling AI operations into OpenRouter's flexible routing layer, Applivo ensures high availability, ultra-low latency, and access to the most capable reasoning models without vendor lock-in.
Configuration
OpenRouter is configured by the system administrator at the server environment level:
- Create an account on openrouter.ai and generate an API key.
- Add your OpenRouter key to your server's .env file.
- Optionally set OPENROUTER_MODEL (defaulting to claude-3-5-sonnet or gpt-4o) and OPENROUTER_SITE_URL.
- Restart the backend service. All AI features activate immediately.
How it works
When you request a cover letter, analyze a job, or run an extension application:
- 1. Context Construction: The backend aggregates the target job requirements, candidate experience graph, and verified achievements.
- 2. Prompt Guardrails: Strict system instructions enforce truthful answers and prohibit fabrication of unearned degrees or skills.
- 3. Model Routing: The request is routed through OpenRouter to the optimal model for the specific task.
- 4. Streaming Delivery: Generated content streams into the frontend interface with instant latency.
- 5. Privacy Enforcement: Zero data retention headers ensure prompt content is not used to train external foundational models.
Available features
- Multi-model routing across verified high-intelligence reasoning architectures.
- Strict prompt guardrails enforcing truthful candidate representations.
- Zero data retention mode for candidate privacy and safety.
- High-speed streaming for instant cover letter drafting and screening answers.
Permissions and data access
Communication with OpenRouter uses server-side API keys with zero data retention enabled. Prompts are processed ephemerally and never persisted by model providers.
Troubleshooting
| Symptom | Likely Cause | Resolution |
|---|---|---|
| AI features return 500 error | OpenRouter server credentials missing or credit balance exhausted | Check API key validity and credit balance on openrouter.ai. |
Frequently asked questions
Are my resumes used to train public AI models?
No. Applivo transmits requests with explicit zero-retention flags, prohibiting model providers from training on candidate data.