# Podar

> Podar is an enterprise AI cost-optimization platform. It sits between your applications and every major AI gateway, optimizes prompts before execution, routes each request to the cheapest model that can answer it well, reuses semantically cached answers, and bills only a percentage of the spend it verifiably removes.

Podar has two engines. Podar Exchange routes each request to the best existing supply source. Podar Foundry manufactures new, smaller supply: when a workload repeats at volume, Podar distills it into a tiny task-specific MicroModel (Podar-Extract, Podar-Classify, Podar-JSON, Podar-SQL and similar), certifies it against a customer-defined quality bar, and routes to it with automatic frontier fallback. Podar distills tasks, not intelligence — it will not build a general-purpose mini model.

Podar ("to prune", Spanish) cuts enterprise AI overspend. In typical enterprise traffic mixes, 40–70% of AI spend is recoverable without any loss of quality. Podar removes that waste rather than reporting on it, and charges roughly 25% of verified realized savings — no upfront license, no per-seat fee.

Key definitions:
- AI Yield: the share of AI spend that bought necessary work at an appropriately sized model. Higher is better.
- AI Wastage: the share spent overpaying for capability the task never needed, or regenerating answers already produced. AI Yield + AI Wastage = 100%.
- Realized savings: per-request difference between the model that would have served a request and the model that did, logged with tokens and quality score, and used as the invoice basis.

Supported gateways: OpenRouter, Portkey, TokenMix, Cloudflare AI Gateway, Kong AI Gateway, LiteLLM, Helicone, Bifrost, plus direct provider connections.

Contact: hello@podar.ai · Product sign-in: https://app.podar.ai

## Pages

- [Home](https://podar.ai/): Overview of Podar, the waste patterns it removes, and the AI Yield metric.
- [How it works](https://podar.ai/how-it-works): The nine stages every request passes through — receive, optimize, score, cache, decompose, route, escalate, validate, measure.
- [Foundry](https://podar.ai/foundry): Podar Foundry distills recurring workloads into tiny task-specific Podar MicroModels and routes to them instead of frontier models.
- [Solutions](https://podar.ai/solutions): Index of the six optimization layers.
- [Pricing](https://podar.ai/pricing): Fee-on-savings model with a worked example on a $7M annual AI bill.
- [About](https://podar.ai/about): Why Podar exists, how it differs from gateways and AI-FinOps tools, and the roadmap.
- [FAQ](https://podar.ai/faq): Answers on savings ranges, pricing, integration effort, quality guarantees, and supported gateways.
- [Contact](https://podar.ai/contact): How to request a free AI spend assessment.

## Solutions

- [Model routing](https://podar.ai/solutions/model-routing): Scores every prompt and sends it to the cheapest model that can answer it well, escalating only when needed.
- [Prompt optimization](https://podar.ai/solutions/prompt-optimization): Removes unnecessary tokens before execution while preserving intent, context, and constraints.
- [Semantic caching](https://podar.ai/solutions/semantic-caching): Reuses answers to semantically equivalent prompts across teams, sessions, and applications.
- [Adaptive workflow routing](https://podar.ai/solutions/adaptive-workflow-routing): Decomposes complex requests into subtasks, routes each to the best-suited model, validates, and recomposes.
- [Multi-gateway fabric](https://podar.ai/solutions/multi-gateway-fabric): Routes across every major AI gateway, comparing price, health, and latency in real time.
- [Podar MicroModels](https://podar.ai/solutions/podar-micromodels): Tiny distilled task-specific models that become new, far cheaper destinations inside the router.
- [AI Yield analytics](https://podar.ai/solutions/ai-yield-analytics): Meters AI Yield and AI Wastage per request with an auditable savings ledger.

## Optional

- [Privacy](https://podar.ai/privacy): How Podar handles request data and telemetry.
- [Terms](https://podar.ai/terms): Website terms of use.