Beta self-hosted HTTP and LLM proxy with Redis-backed caching, non-streaming LLM idempotency, and budget guardrails.
-
Updated
Oct 12, 2026 - Python
Beta self-hosted HTTP and LLM proxy with Redis-backed caching, non-streaming LLM idempotency, and budget guardrails.
Open protocol for runtime budget and action authority over AI agents
Runaway AI agent spend prevention — demo using Cycles ($6 in 30s → hard stop at $1)
Observability and local budget control for x402 AI-agent payments — Python port of @harpd/observe.
Policy & authorization layer for AI agent payments: ALLOW / APPROVAL / DENY with budgets, merchant lists, signed one-time grants and tamper-evident audit. Python + MCP.
Custos - Intelligent LLM Gateway with smart routing, caching, and cost governance
Production-grade guardrails for AI agent function calls. Budget, scope, rate-limit, PII & deletion guards with Human-in-the-Loop approval.
AI agent security SDK for runtime guardrails: control tool calls, source-chain provenance, budget limits, plan preflight, prompt-injection isolation, and adjudication tokens. Apache-2.0 Python SDK.
Execution safety & governance layer for AI agents: idempotency, budgets, tool control, audit logs.
Deterministic run-scoped governance layer for local coding agents.
Python SDK for pre-call AI budget controls and customer-, feature-, and workflow-level cost attribution.
Single-host bounded-concurrency Agent DAG scheduler for DSH Surge Mode, with fail-closed model adapters and a JSON task-plan CLI.
Safety layer for AI agents — Protect · Route · Prove. Budgets, tool-loop hard stops, failover, chaos DR proof. OpenAI drop-in · MCP · n8n.
AI agent action authority — demo of blocking unapproved tool calls with Cycles
Typed Python tenant onboarding with an Infrai monthly spend gate
Flight recorder + hard spend caps for AI coding agents (Claude Code, pi). Halts the run at max_run_cost or a per-agent max_cost. Local SQLite traces, replay UI. MIT.
Python order workflow with an Infrai monthly hard cap
Multi-agent orchestration on MCP: a state server, six-phase dependency graph, approval gates, and budget control.
💸 budget-pacer —— 面向 MaiBot 的月度 LLM 预算节奏闭环调节插件。设定月度 API 预算后自动动态调整 Bot 发言频率:花费超前于时间进度则降频,花费滞后可加频,让整月累计花费沿预算斜率平稳推进;支持单群例外额度、死区防抖、指数平滑、超支硬闸门,跨月自动复位,全程无需人工干预。
To associate your repository with the budget-control topic, visit your repo's landing page and select "manage topics."