AI agent infrastructure tools
66 agent infrastructure tools in the AgentScout directory, each with a link-checked listing. The infrastructure behind AI agents that is not primarily MCP: LLM and AI gateways, inference clouds, agent runtimes and control planes, tool-integration and auth platforms, agent email, payments and identity layers, and agent filesystems: LiteLLM, Portkey, Composio, Nango, Groq, Replicate, Together AI and more. MCP servers have their own page.
Updated
There are 66 agent infrastructure tools: 41 open source, 1 source-available, 11 free or freemium, 11 paid, usage-based or enterprise, 2 pricing on the maker's site.
Quick comparison Top 10 of 66 listed below. Pricing as the maker states it; blank means we don't have it.
| Tool | Pricing | Best for | Open source |
|---|---|---|---|
 LayerFS | Open source | Yes | |
 OpenComputer | Usage-based | Build an agent | |
 HiRoute | Open source | Write code | Yes |
 Sentio | Open source | Automate workflows | Yes |
 Agent Tincan | Open source | Build multi-agent systems | Yes |
 Agent Barn | Open source + cloud | Monitor & evaluate agents | Yes |
 opencodex | Open source | Write code | Yes |
 Weave Router | Open source | Write code | Yes |
 LiteLLM Agent Platform | Open source | Yes | |
 Caspian | Open source | Build an agent | Yes |
Tasks in this category
Showing 66 of 66 tools
LayerFS
LayerFS is a filesystem built for running many coding agents in parallel. Each agent mounts a complete, writable workspace on top of shared committed history, does its work, commits what it changed and goes away, so compute only exists while a command runs. Every tool call becomes a commit on a branch, same-file clashes between agents are refused rather than silently merged, each line records the step that wrote it, and any commit can be forked at no cost for recursive exploration or RL training. Agents on different computers share one serverless store with chunk-level deduplication.
Added
OpenComputer
OpenComputer is a platform for building and deploying long-running AI agents as code. You define an agent as a TypeScript function with its model, tools and MCP servers, deploy it with one command, and every session gets its own real cloud computer, with managed sessions, schedules and API keys kept outside the agent's runtime. Templates include a coding agent that edits a repo, runs checks and opens a draft pull request, and market-research and GTM agents; its own Slack coding agent runs on it. Pricing is pay-as-you-go for model calls and per-second machine time.
Added
HiRoute
HiRoute, from the Higress team, is an open-source local model router for long-running agents like Codex and Claude Code: it hands routine stages to cheaper models, saves your strongest model for hard steps and keeps a record of each decision.
Added
Sentio
Sentio, from Truespar, is a self-hosted email inbox API for AI agents: a Rust SMTP server that gives each agent a real address, delivers inbound mail as authenticated webhooks and lets agents reply in-thread over REST, with multi-tenant isolation.
Added
Agent Tincan
Agent Tincan lets your AI agents (Claude Code, Codex, Gemini, Hermes, ChatGPT and others) ask each other to do things and hand back results through a small relay on your own Tailscale network, with no open ports or API keys between agents.
Added
Agent Barn
Agent Barn is an open-source (Apache-2.0) control plane for running a fleet of AI agents on your own Kubernetes: create, configure, run and monitor agents across Slack, Microsoft Teams and Telegram from one place.
Added
opencodex
A local provider proxy that lets OpenAI Codex (CLI, app and SDK) and Claude Code run on other model providers such as Claude, Gemini, Grok, DeepSeek or local Ollama models.
Added
Weave Router
An open-source LLM router for Claude Code, Codex and other coding agents that sends each request to the cheapest model that still meets your quality bar, cutting model spend.
Added
LiteLLM Agent Platform
A control plane from the LiteLLM team that gives one API and gateway for calling many agents, such as OpenCode, Claude Managed Agents, Cursor Agents and Deep Agents.
Added
Caspian
An open-source communication SDK that lets AI agents send and receive messages over email, WhatsApp, Slack, Discord and other channels.
Added
TypeSafe Jev
Jev is TypeSafe's System One decision model, built for the small decisions inside software and agents. You send a state and typed questions (choice, score and more) and get typed answers with probabilities back instead of free-form text, so code can branch on them directly to route, classify, score or rank options. Agents can use it to settle narrow calls cheaply and save the larger LLM for open-ended work. Available through TypeSafe's API, Python and JavaScript SDKs, an agent skill, and OpenRouter.
Added
Grok Bot Jev Router
Grok Bot Jev Router is an open-source Python reference router that puts TypeSafe Jev in front of Grok Bot as a cheap decision layer. Before expensive research, browser, retry or subagent work, Jev classifies the request and returns an explicit action such as reuse a cached result, stop a failing retry, cap research, run a deterministic step, or ask a human. It starts in shadow mode that only logs decisions, has an active mode for a Grok Bot skill that honors them, and includes a kill switch. It never sends, pays or deletes anything itself.
Added
Monid
Gateway that gives an agent one base URL and one key to call thousands of paid tools, from web scraping and enrichment to image and voice generation, billed per call.
Added
Treg
Single endpoint and token that lets agents call a catalog of third-party APIs for SEO, social data, enrichment, scraping and media generation, priced per call.
Added
HarnessRouter
One API for running coding-agent harnesses such as Codex, Claude Code and Hermes as product backends, with sessions, streaming, files and failure handling.
Added
Agent Mesh
Private peer-to-peer network for AI agents: each node publishes tools and models, finds other agents' tools and carries a verifiable identity.
Added
Magpie
Free, open-source menu bar app and local gateway that lets any coding agent (Claude Code, Codex, Gemini CLI, OpenCode and 60+ more) use any model provider or subscription, with failover when a quota runs out.
Added
Vectoree
Vectoree gives coding agents a backend behind one project key: an OpenAI-compatible model gateway across providers, an MCP tool hub, auth, Postgres and storage, driven by a skill and CLI, pay-as-you-go with no monthly fee.
Added
XCrawl
XCrawl is a web scraping, crawling and SERP API that returns clean JSON, Markdown or screenshots in one request, with agent skills for plugging it into AI agents; free trial with no card.
Added
zero.xyz
Lets AI agents discover, connect to and pay per call for thousands of APIs and services without API-key setup or separate sign-ups.
Added
Kong Gateway
Cloud-native API, LLM and MCP gateway that adds routing, auth, multi-LLM support, semantic security and MCP traffic analytics through plugins.
Added
Novu
Open-source communication infrastructure that gives products and AI agents one API to reach users over in-app inbox, email, SMS, push, Slack, Teams and chat.
Added
Portkey AI Gateway
Fast open-source AI gateway that routes requests to hundreds of LLMs through one API with fallbacks, load balancing and built-in guardrails.
Added
SkillHub
Self-hosted registry for publishing, versioning and governing agent skill packages across an organization, with RBAC and audit logs.
Added
Open Terminal
Self-hosted terminal service from the Open WebUI team that gives AI agents an API to run commands, manage files and execute code.
Added
EigenFlux
Open-source communication and broadcast network where AI agents share information, needs and capabilities.
Added
LiteLLM
Open-source AI gateway and Python SDK that calls 100+ LLM providers in the OpenAI format, with cost tracking, guardrails, load balancing and logging.
Added
Bifrost
High-performance AI gateway from Maxim that unifies 20+ LLM providers behind one OpenAI-compatible API and acts as an MCP and agent gateway.
Added
Manifest
Open-source LLM gateway that routes each request from agents and apps to the right model across API keys, subscriptions and local models, with spend tracking.
Added
Omnara
Open-source managed agent platform, a self-hostable alternative to Claude Managed Agents and the OpenAI Agents API, with tools, secrets and sandboxes.
Added
Atomic Mail
Email API that gives AI agents their own inboxes, built on JMAP with MCP and agent skill support.
Added
OmniRoute
Self-hosted AI gateway that puts hundreds of model providers behind one endpoint for Claude Code, Codex, Cursor and others, with quota-aware fallback and MCP/A2A support.
Added
Tyk Gateway
Cloud-native API and AI gateway supporting REST, GraphQL, gRPC and MCP, with rate limiting, auth and analytics.
Added
Arcade
Actions runtime for enterprise agents with thousands of prebuilt authenticated tools, plus an open-source framework for building your own MCP servers.
Added
Nango
Integration platform that connects products and AI agents to 1,000+ APIs with managed auth, syncs and tool calls.
Added
CLI-Anything
Open-source project from HKUDS that generates agent-native command-line interfaces for existing software, codebases and web APIs, with CLI-Hub, a registry of community CLIs that coding agents can install and drive.
Added
skills.sh
Vercel Labs' open-source CLI (npx skills) and directory for finding, installing and sharing agent skills across Claude Code, Codex, Cursor, OpenCode and 75+ other coding agents.
Added
Agent Resource Market
Eellos LLC's API-first beta where AI agents discover and publish resource offers, join structured discussions and post contextual reviews. Exchanges are simulated with demo credits (payments are disabled) and registration is capacity-limited.
Added
NVIDIA Open Agent Safety Platform
Reference stack pairing OpenShell with NVIDIA Sentry on BlueField-4 DPUs for hardware-level agent monitoring.
Added
Composio
Tool integrations and auth for agents across hundreds of apps.
Added
Self Agent Vault
Payment and login layer for AI agents: agents check out and sign in using mock tokens, while real card numbers and passwords are swapped in out of the agent's view, with spending caps, merchant allowlists and a kill switch.
Added
Groq
Fast inference cloud running open models on custom LPU chips, used for low-latency agents.
Added
Recall.ai
API for meeting bots that join Zoom, Meet and Teams to record and transcribe.
Added
Replicate
Run and deploy open models through an API, with Cog for packaging models.
Added
Together AI
AI cloud for running, fine-tuning and serving open models.
Added
TollBit
Lets publishers charge AI agents and bots for accessing their content.
Added
Google Workspace CLI
Official Google Workspace command-line tool for Drive, Gmail, Calendar, Docs and more, usable by agents.
Added
Corsair
Integration layer for connecting your users' apps to agents.
Added
Casdoor
Agent-first identity and access management with an MCP and agent gateway.
Added
CocoIndex
Incremental data engine for keeping agent context and indexes fresh.
Added
Electric
Sync engine and agent platform that keeps data in sync between Postgres, apps and agents in real time.
Added
Bindu
Identity, communication and payments layer that lets AI agents talk to and pay each other.
Added
Nasiko
Open runtime for deploying and operating AI agents.
Added
Reef
Infrastructure for continually self-improving agents.
Added
Open Connector
Open-source auth gateway connecting SaaS apps to AI agents via SDK, CLI and MCP.
Added
Olares
Open-source personal cloud OS for running always-on agents at home.
Added
ACI.dev
Open-source tool-calling platform connecting hundreds of tools to agents.
Added
Liveblocks
Realtime collaboration infrastructure for multiplayer apps and AI agents.
Added
optillm
Inference proxy that applies reasoning techniques to improve LLM answers.
Added
Mirage
Virtual terminal environment for AI agents.
Added
OneCLI
Sandboxed agent harness for teams with secure credential handling.
Added
AgentFS
Turso's filesystem for agents with snapshots and auditing.
Added
Ref
Token-efficient documentation search for coding agents.
Added
Agentcard
Gives each user of your product a payment card wallet that their AI agent can spend from within set limits.
Added
Aident AI
Connects AI agents to more than 1,000 apps through one integration so they can act across tools.
Added
Context.dev
Web data API for AI agents that scrapes pages to Markdown, searches the web and crawls sites.
Added
Top agent infra repos on GitHub (52)
| Repo | GitHub stars | Licence | Last push |
|---|---|---|---|
| OmniRoute ↗ | 77,863 | MIT | 2026-10-21 |
| LiteLLM ↗ | 61,465 | MIT | 2026-10-21 |
| CLI-Anything ↗ | 54,399 | Apache License 2.0 | 2026-09-22 |
| Kong Gateway ↗ | 44,278 | Apache License 2.0 | 2026-10-19 |
| Novu ↗ | 40,164 | — | 2026-10-21 |
| skills.sh ↗ | 35,459 | MIT | 2026-10-19 |
| Google Workspace CLI ↗ | 31,557 | Apache License 2.0 | 2026-03-31 |
| opencodex ↗ | 17,682 | — | 2026-10-21 |
| Magpie ↗ | 16,857 | MIT | 2026-10-21 |
| Corsair ↗ | 14,922 | Apache-2.0 | 2026-10-20 |
| Casdoor ↗ | 14,592 | Apache License 2.0 | 2026-10-21 |
| Portkey AI Gateway ↗ | 13,214 | MIT | 2026-05-25 |
| Nango ↗ | 12,807 | Elastic License 2.0 (source-available) | 2026-10-21 |
| CocoIndex ↗ | 11,710 | Apache License 2.0 | 2026-10-21 |
| Tyk Gateway ↗ | 10,863 | Mozilla Public License 2.0 (except the ee folder) | 2026-10-21 |
Top 15 of 52 by stars. Stars and licence come from each repo, checked when the listing was last verified.
All categories
Who uses these
Best-of lists
Alternatives pages
Free tools for this: Agent run cost estimator · Build vs buy calculator
Facts come from our directory listings; pricing is as each maker states it. Check official sites before you choose.