VERIFIED
MIDDLEWARE FOR
AI-AGENT TRAFFIC.

Each module declares which providers it works with, which provider capabilities it requires, how it falls back when a capability is missing, and which receipt fields it stamps. Drop them into your PRXY_PIPE; the receipt's module_chain is the audit trail.

12 official · 0 community · 12 verified · works with Anthropic, OpenAI, Google, Groq, Bedrock, OpenRouter · cloud + local. Verified-publisher program open.

Allcachecontextinjectionoptimizationroutingsafety✓ Verified only

Featured

curated by the prxy team
Compaction Bridge✓ Verifiedproduction
@prxy-official/compaction-bridge
Survives Claude Code's auto-compaction without state loss. Detects post-compaction requests and re-injects the most relevant pieces of the most recent eviction archive into the system prompt.
contextMITcloud yeslocal yes
IPC — Inter-Prompt Compression✓ Verifiedproduction
@prxy-official/ipc
Compresses older messages instead of dropping them when a request exceeds the context window utilization target. Keeps long conversations going past context limits.
contextMITcloud yeslocal yes
MCP Optimizer✓ Verifiedproduction
@prxy-official/mcp-optimizer
Prunes MCP tools to the ones actually relevant to the user's last message. Solves the 67k-tokens-of-MCP-tools problem by scoring each tool against the request via embeddings.
optimizationMITcloud yeslocal yes
Patterns — Golden Loop✓ Verifiedproduction
@prxy-official/patterns
Injects relevant past patterns into the system prompt and forges new ones from successful resolutions. The Golden Loop — your gateway gets smarter over time.
injectionMITcloud yeslocal yes
Semantic Cache✓ Verifiedproduction
@prxy-official/semantic-cache
Returns cached responses for semantically similar (not just identical) requests. Embeds the request, vector-searches past requests, replays the cached response if similarity is above threshold.
cacheMITcloud yeslocal yes

cache

view all →
Exact Cache✓ Verifiedproduction
@prxy-official/exact-cache
Returns cached responses for byte-identical requests. Hash-keyed by a stable serialization of the canonical request — fastest cache type with zero false positives.
cacheMITcloud yeslocal yes
Semantic Cache✓ Verifiedproduction
@prxy-official/semantic-cache
Returns cached responses for semantically similar (not just identical) requests. Embeds the request, vector-searches past requests, replays the cached response if similarity is above threshold.
cacheMITcloud yeslocal yes
Tool Cache✓ Verifiedpreview
@prxy-official/tool-cache
Observes (and in v2 will cache) MCP tool-call results. Records tool_use → tool_result pairs and stamps observability metadata for repeated calls. Excludes side-effecting tools by default.
cacheMITcloud yeslocal yes

context

view all →
Compaction Bridge✓ Verifiedproduction
@prxy-official/compaction-bridge
Survives Claude Code's auto-compaction without state loss. Detects post-compaction requests and re-injects the most relevant pieces of the most recent eviction archive into the system prompt.
contextMITcloud yeslocal yes
IPC — Inter-Prompt Compression✓ Verifiedproduction
@prxy-official/ipc
Compresses older messages instead of dropping them when a request exceeds the context window utilization target. Keeps long conversations going past context limits.
contextMITcloud yeslocal yes
Rehydrator✓ Verifiedproduction
@prxy-official/rehydrator
Pulls archived context back when the user references past turns. Companion to ipc — when the user says 'remember when…', searches eviction archives semantically and re-injects the most relevant turns.
contextMITcloud yeslocal yes

injection

view all →
Patterns — Golden Loop✓ Verifiedproduction
@prxy-official/patterns
Injects relevant past patterns into the system prompt and forges new ones from successful resolutions. The Golden Loop — your gateway gets smarter over time.
injectionMITcloud yeslocal yes

optimization

view all →
MCP Optimizer✓ Verifiedproduction
@prxy-official/mcp-optimizer
Prunes MCP tools to the ones actually relevant to the user's last message. Solves the 67k-tokens-of-MCP-tools problem by scoring each tool against the request via embeddings.
optimizationMITcloud yeslocal yes
Prompt Optimizer✓ Verifiedproduction
@prxy-official/prompt-optimizer
Restructures requests to maximize Anthropic prompt-cache hits. Sorts tools deterministically and stamps cache-control markers on the longest stable prefix.
optimizationMITcloud yeslocal yes

routing

view all →
Router✓ Verifiedproduction
@prxy-official/router
Picks the best model for each request. Strategies: cheapest-first, fallback chain, or q-learning based on per-query historical success rate.
routingMITcloud yeslocal yes

safety

view all →
Cost Guard✓ Verifiedproduction
@prxy-official/cost-guard
Per-request, per-day, and per-month USD budget enforcement. Short-circuits requests that would breach a configured limit with a 429-shaped response.
safetyMITcloud yeslocal partial
Guardrails✓ Verifiedproduction
@prxy-official/guardrails
Content filtering at the gateway layer. PII redaction (email, SSN, credit-card), profanity blocking, and user-supplied regex patterns. v2 adds NIM and Anthropic Constitutional backends.
safetyMITcloud yeslocal yes

First 100 verified publishers

community modules · free today · paid modules planned

Build a module with @prxy/module-sdk, publish it to npm under your scope, then submit a PR adding a manifest to the public registry. The first 100 accepted community publishers get priority verification review and launch-page featuring. Paid-module revenue share is planned, not live; terms will be published before paid modules are accepted.

Read the publishing guide →