Paperclip Alternatives and Competitors, by Category

The most common mistake when evaluating AI agent tools is treating them as a single category. The products that get compared to Paperclip span four distinct layers of the stack, and each layer answers a different question. Comparing Paperclip to LangChain is structurally like comparing a company to a recruiting firm's job-description template: related in subject matter, different in kind.

This guide sorts the field before comparing. The category question comes first.

Orchestrators

An orchestrator manages a collection of agents as an organisation: goals, reporting lines, budgets, approval gates, and a mechanism for recovery when something goes wrong.

Paperclip is in this category. It is an open-source platform for running teams of AI agents. The relationship it describes is explicit in its own documentation: "If OpenClaw is an employee, Paperclip is the company." The layer Paperclip occupies is the org, not the individual agent. It provides an org chart, governance rules, agent mandates, and a ticketing system for the work. The agents it manages run on runtimes that live at a layer below.

No serious direct competitor occupies the same position: open-source, self-hostable, infrastructure-owned, with a working orchestration layer already included. The closest comparison points are either in a different category or at a different layer of the stack.

Agent runtimes

A runtime is what an individual agent actually runs on: its memory, toolset, skill library, and model access. Runtimes and orchestrators are complementary layers. A runtime without an orchestrator runs a single agent in isolation. An orchestrator without a runtime specification defers the capability question to whichever runtime it selects per agent.

Hermes, built by Nous Research, is the runtime most directly associated with Paperclip. It carries persistent memory across sessions, 30+ native tools, 80+ skills, and multi-provider model routing across eight inference providers. Hermes runs as a built-in Paperclip adapter type: hermes_local for agents on the same machine, hermes_gateway for remote instances. The orchestrator-versus-runtime distinction, and when to run Hermes inside Paperclip versus standalone, is covered in full in Paperclip vs Hermes.

OpenClaw is an open-source personal AI assistant you run on your own devices. It is MIT-licensed, runs as a self-hosted Node.js gateway, and responds on the channels you already use. It composes with Paperclip rather than competing with it at the orchestrator layer: Paperclip's adapter catalogue includes openclaw_gateway, which lets a Paperclip company deploy an OpenClaw instance as a managed employee in its org chart. As a standalone product, OpenClaw is a personal assistant runtime. It does not manage teams of other agents.

Build-your-own frameworks

LangChain, LangGraph, CrewAI, and Microsoft Agent Framework are developer libraries. They provide building blocks for agent logic; the developer writes the orchestration.

LangChain is the most widely adopted open-source framework for building LLM applications. It handles single-agent logic and tool use as a coding library.

LangGraph is a separate, lower-level project from the LangChain team for stateful multi-agent systems. It handles loops, persistence, and graph-based execution flows. The developer writes the control flow in code; LangGraph provides the state management.

CrewAI models agent teams with defined roles. It is self-contained with no external framework dependencies and is oriented toward structured, role-based multi-agent pipelines.

Microsoft Agent Framework is Microsoft's current investment in agentic developer tooling, positioned as the successor to AutoGen and Semantic Kernel, both of which are now in maintenance mode.

The structural distinction from Paperclip is this: these frameworks give you the raw material to build your own orchestration layer. Paperclip is an orchestration layer you deploy. A team building on LangGraph is writing its own Paperclip. A team running Paperclip is operating one.

Managed and hosted clouds

The hosted cloud platforms offer fully managed agent infrastructure. The infrastructure runs on the vendor's systems; you configure and deploy agents through their console.

Microsoft Copilot Studio is hosted on Azure infrastructure. Its documentation specifies deployment to Azure datacenters by geographic region.

Gemini Enterprise Agent Platform is a Google Cloud service rebranded from Vertex AI Agent Builder at Cloud Next 26 in April 2026. Existing customers required no migration. It is a fully managed service running on Google infrastructure.

Amazon Bedrock AgentCore is AWS's fully managed agent infrastructure layer. It handles memory, execution environment, and tool connectivity without server management.

Salesforce Agentforce runs on Salesforce and AWS Hyperforce infrastructure. It is oriented toward CRM-adjacent workflows and requires Salesforce Data Cloud for advanced orchestration features.

IBM watsonx Orchestrate is available as SaaS on IBM Cloud and on AWS. IBM is the exception in this group: it also offers an on-premises deployment path at the enterprise tier. The cloud SaaS offering is the default; on-prem is available separately.

The consistent trade-off across the hosted category is data custody. Agents run on vendor infrastructure, and data is processed there. There is no self-hosted or infrastructure-owned deployment path with any of these products, except the enterprise on-prem tier IBM makes available for watsonx.

How the categories bear on the choice

The choice between categories is a governance and data-sovereignty question you answer before the product comparison starts.

If you need to write your own orchestration logic, the build-your-own frameworks are the relevant category. For organisations already running on Azure, Google Cloud, AWS, or Salesforce where vendor data custody is acceptable, the managed clouds remove operational overhead.

Paperclip fits the remaining case: a working orchestration layer on your own infrastructure, with data staying local and no dependency on a vendor's managed service. Its installation documentation is direct: all data stays local, and nothing is sent to an external server beyond the API calls your agents make to model providers. That guarantee is structurally incompatible with the managed cloud model.

The runtime question is separate. Once the orchestrator layer is chosen, the decision about which runtime capabilities each agent needs, including persistent memory and model-agnostic routing, is a distinct evaluation. For the Hermes side of that, Paperclip vs Hermes covers it in full.

Get on the list for the next piece in the series.