Use Codex at 35.6% lower costs
Edgee compresses prompts before they reach LLM providers and reduces token costs by up to 50%. Same code, fewer tokens, lower bills.
Edgee compresses input and output tokens by up to 50% through surgical removal of redundancy (tool-result trimming, tool-surface reduction, output brevity). Works as a transparent proxy—install CLI, connect coding agent, save tokens on first request. No application code changes required.
Edgee provides automatic per-request retry and fallback across 200+ models from 8+ providers (OpenAI, Anthropic, Google, Mistral, DeepSeek, xAI, zAI, AWS Bedrock). Transient errors retry once; persistent failures route to configured backup. Zero downtime from agent perspective.
Bring Your Own Keys (BYOK) feature lets users register API keys for Anthropic, OpenAI, Google Vertex AI, Mistral, DeepSeek, xAI, zAI, and AWS Bedrock. Requests route through user's key and billing account while Edgee's compression, fallback, and observability continue to work.
Edgee implements plan-cap continuity: when Claude Pro/Max quota is exhausted, it automatically falls back to API-key-based providers, keeping the session active. Compression also extends sessions by 26.5% on the same plan.
Edgee dashboard aggregates costs, token usage, compression savings, and errors across all requests. Team plan includes per-repo/per-PR attribution via GitHub integration, spending caps, alerts, and usage exports. Every response includes real-time compression metrics (saved tokens, cost savings, reduction percentage).