114 lines
2.4 KiB
Markdown
114 lines
2.4 KiB
Markdown
# Real-World Examples
|
|
|
|
Production scenarios demonstrating VAPORA solving business problems.
|
|
|
|
## Use Cases
|
|
|
|
### 01: Code Review Pipeline
|
|
**File**: `01-code-review-workflow.rs`
|
|
|
|
Automated multi-agent code review with cost optimization.
|
|
|
|
**Business Value**:
|
|
- Review 50 PRs/day consistently
|
|
- 99% accuracy matching human review
|
|
- $500+/month savings vs manual review
|
|
- Instant feedback to developers
|
|
|
|
**Architecture**:
|
|
1. Ollama (free) - Static analysis
|
|
2. GPT-4 ($10/1M) - Quality review
|
|
3. Claude ($15/1M) - Architecture review
|
|
|
|
**Run**:
|
|
```bash
|
|
cargo run --example 01-code-review-workflow
|
|
```
|
|
|
|
### 02: Documentation Generation
|
|
**File**: `02-documentation-generation.rs`
|
|
|
|
Automatically generate and maintain API documentation.
|
|
|
|
**Business Value**:
|
|
- Docs always in sync with code
|
|
- 2-3 week lag → instant updates
|
|
- 99%+ accuracy vs manual docs
|
|
- $1000+/month savings
|
|
|
|
**Architecture**:
|
|
1. Ollama - Code analysis
|
|
2. Claude - Doc writing
|
|
3. GPT-4 - Quality check
|
|
|
|
**Run**:
|
|
```bash
|
|
cargo run --example 02-documentation-generation
|
|
```
|
|
|
|
### 03: Issue Triage
|
|
**File**: `03-issue-triage.rs`
|
|
|
|
AI-powered issue classification and routing.
|
|
|
|
**Business Value**:
|
|
- 200+ issues/month triaged instantly
|
|
- Consistent classification criteria
|
|
- 95% accuracy
|
|
- $1000+/month savings (vs 20 hours manual work)
|
|
|
|
**Architecture**:
|
|
1. Ollama (free) - Initial classification
|
|
2. Claude ($0.08) - Detailed analysis when needed
|
|
|
|
**Run**:
|
|
```bash
|
|
cargo run --example 03-issue-triage
|
|
```
|
|
|
|
## Cost Analysis Pattern
|
|
|
|
Real-world examples use **cost-aware multi-stage pipelines**:
|
|
|
|
```
|
|
Stage 1 (Cheap): Ollama/GPT-4 for routing
|
|
↓
|
|
Stage 2 (Premium): Claude only when needed
|
|
↓
|
|
Stage 3 (Validation): GPT-4 for quality check
|
|
```
|
|
|
|
This achieves:
|
|
- 95%+ accuracy at 10% of premium cost
|
|
- Fast iteration with feedback
|
|
- Controlled budget
|
|
|
|
## Key Metrics
|
|
|
|
Each example demonstrates:
|
|
|
|
| Metric | Typical Value |
|
|
|--------|--------------|
|
|
| Accuracy | 95%+ vs manual |
|
|
| Speed | 10-100x faster |
|
|
| Cost | 1-10% of manual |
|
|
| Coverage | 100% automated |
|
|
|
|
## Adapting for Your Use Case
|
|
|
|
1. **Choose pipeline stages**: Cheap → Medium → Premium
|
|
2. **Adjust triggers**: When to escalate to next stage
|
|
3. **Set budgets**: Per-role monthly limits
|
|
4. **Monitor quality**: Track accuracy vs manual baseline
|
|
|
|
## Next Steps
|
|
|
|
- Implement case-specific routing rules
|
|
- Integrate with your systems (GitHub, Jira, etc.)
|
|
- Monitor and adjust thresholds
|
|
- Expand to additional use cases
|
|
|
|
---
|
|
|
|
Navigate to `../` for more examples.
|