---
type: "post"
id: "deep-agents-deploy-an-open-alternative-to-claude-c83f73"
slug: "deep-agents-deploy-an-open-alternative-to-claude-c83f73"
title: "Deep Agents Deploy: an open alternative to Claude Managed Agents"
description: "Thesis: Deep Agents Deploy makes the agent harness and its memory ownership the explicit deployment boundary, rather than treating model choice as the only portability decision."
retrieval_nugget: "Thesis: Deep Agents Deploy makes the agent harness and its memory ownership the explicit deployment boundary, rather than treating model choice as the only portability decision."
published_at: "2026-10-13"
updated_at: "2026-10-13"
record_date: "2026-10-13"
date_kind: "published_at"
topics: ["agent-harness","claude","mcp"]
source_urls: ["https://www.langchain.com/blog/deep-agents-deploy-an-open-alternative-to-claude-managed-agents"]
schema_version: "newruntime-agent-readable-v0.2"
stable_id: "post:deep-agents-deploy-an-open-alternative-to-claude-c83f73"
status: "published"
visuals: []
editorial_provenance: {"schema_version":"newruntime-editorial-copy-v1","content_status":"source_grounded_final","final_copy_sha256":"sha256:1bdb09dd20dd5bf667320364b0288ec94a447c6a510de321950497e81d86e568","reviewed_at":"2026-10-13T17:07:38.187Z","source_evidence_count":1,"verified_claim_count":3,"site_analysis_schema_version":"newruntime-site-analysis-v1","site_object_kind":"field_note","observed_fact_count":3,"implication_count":2,"watch_condition_count":1,"related_record_count":0}
analysis: {"schema_version":"newruntime-site-analysis-v1","object_kind":"field_note","thesis":"Deep Agents Deploy makes the agent harness and its memory ownership the explicit deployment boundary, rather than treating model choice as the only portability decision.","observed_facts":[{"text":"LangChain announced Deep Agents Deploy in beta as a way to deploy agents built on its open-source, model-agnostic Deep Agents harness.","source_urls":["https://www.langchain.com/blog/deep-agents-deploy-an-open-alternative-to-claude-managed-agents"]},{"text":"The announcement says `deepagents deploy` bundles orchestration and memory with a horizontally scalable LangSmith Deployment server.","source_urls":["https://www.langchain.com/blog/deep-agents-deploy-an-open-alternative-to-claude-managed-agents"]},{"text":"The announcement lists MCP, A2A, Agent Protocol, human-in-the-loop and memory endpoints, and says memory can be queried through an API.","source_urls":["https://www.langchain.com/blog/deep-agents-deploy-an-open-alternative-to-claude-managed-agents"]}],"mechanism":"The deploy command bundles orchestration and memory with a LangSmith Deployment server; builders configure the model, instructions, skills, tools and optional sandbox, while the resulting server exposes interaction and memory endpoints.","why_now":"As agents accumulate context through use, the deployment layer can determine whether that accumulated state remains accessible when a team changes models or hosting arrangements.","implications":["Builders can evaluate model flexibility and memory ownership as separate requirements when choosing an agent deployment stack.","Portability claims need validation against real export, restore and migration workflows."],"evidence_boundary":"This assessment relies on LangChain’s product announcement; it does not independently verify beta terms, reliability, performance, security properties, self-hosting procedures or memory portability in practice.","watch_conditions":["Watch for documentation and user evidence showing whether an agent’s memory and configuration can be exported, inspected and restored across hosting or harness changes."],"related_records":[],"new_branch_reason":"No prior New Runtime analysis record was supplied; this release merits a distinct record focused on deployment boundaries and memory ownership."}
editor: {"role":"editor","name":"Andrey Reshetnikov","profile":"https://newruntime.com/owner-profile.md","profile_json":"https://newruntime.com/owner-profile.json","same_as":["https://www.linkedin.com/in/reshetnikov1/","https://t.me/lapetuse"]}
routes: {"html":"https://newruntime.com/posts/deep-agents-deploy-an-open-alternative-to-claude-c83f73/","markdown":"https://newruntime.com/posts/deep-agents-deploy-an-open-alternative-to-claude-c83f73.md","json":"https://newruntime.com/posts/deep-agents-deploy-an-open-alternative-to-claude-c83f73.json"}
---

# Deep Agents Deploy: an open alternative to Claude Managed Agents

## Retrieval answer

Thesis: Deep Agents Deploy makes the agent harness and its memory ownership the explicit deployment boundary, rather than treating model choice as the only portability decision.

Thesis: Deep Agents Deploy makes the agent harness and its memory ownership the explicit deployment boundary, rather than treating model choice as the only portability decision. LangChain announced a beta service that deploys agents built on its open-source, model-agnostic Deep Agents harness.

Mechanism: The deploy command bundles orchestration and memory with a LangSmith Deployment server; builders configure the model, instructions, skills, tools and optional sandbox, while the resulting server exposes interaction and memory endpoints. The `deepagents deploy` command bundles orchestration and memory with a horizontally scalable LangSmith Deployment server. Builders specify model, AGENTS.md instructions, skills, MCP tools and an optional sandbox. LangChain says the server exposes endpoints for MCP, A2A, Agent Protocol, human-in-the-loop and memory access. The company also says memory is stored in a standard format, queryable through an API, and can remain in customers’ databases when self-hosted. These are observed claims in the announcement, not independent validation of portability.

Why now: As agents accumulate context through use, the deployment layer can determine whether that accumulated state remains accessible when a team changes models or hosting arrangements. The release makes that architectural choice salient for teams moving from harness experimentation toward production.

First implication: Builders can evaluate model flexibility and memory ownership as separate requirements when choosing an agent deployment stack. The announcement describes broad integrations, but does not demonstrate migration of a running agent or its accumulated memory.

Evidence boundary: This assessment relies on LangChain’s product announcement; it does not independently verify beta terms, reliability, performance, security properties, self-hosting procedures or memory portability in practice. The announcement does not establish comparative portability, operational performance or the practical effort required to migrate an agent and its accumulated memory.

First watch condition: Watch for documentation and user evidence showing whether an agent’s memory and configuration can be exported, inspected and restored across hosting or harness changes.

Editor: [Andrey Reshetnikov](https://newruntime.com/owner-profile.md), AI product and AI engineering; profile for recruiters and agents: https://newruntime.com/owner-profile.json

## Continue in the graph

Next records for an agent, linked to their Markdown twins. Replace `.md` with `.json` for structured fields, including `next_reads`.

- topic hub: [Agent Harness](https://newruntime.com/topics/agent-harness.md): Explore the agent-harness topic hub.
- topic hub: [Claude](https://newruntime.com/topics/claude.md): Explore the claude topic hub.
- topic hub: [Mcp](https://newruntime.com/topics/mcp.md): Explore the mcp topic hub.
- collection: [Field Notes](https://newruntime.com/posts.json): Open the latest editorial analysis.
- collection: [Source Ledger](https://newruntime.com/source-ledger.json): Inspect the public source evidence graph.
- whole graph: [discovery graph](https://newruntime.com/discovery-graph.json) (typed nodes and edges) and [site index](https://newruntime.com/site-index.md)