update the comparison page#2770
Conversation
|
Mintlify preview ID generated: preview-naomid-1771859602-4cee604 |
|
Mintlify preview ID generated: preview-naomid-1771863269-606c7c9 |
|
Mintlify preview ID generated: preview-naomid-1771870632-3147185 |
|
Mintlify preview ID generated: preview-naomid-1771871114-3f4b6bb |
| | ------ | ------------------------- | -------------------- | ------------- | | ||
| | **Primary use case** | Interactive coding agent in terminal or IDE; Build production agents programmatically with SDK | Build production agents programmatically | Interactive coding agent (CLI, desktop, IDE, cloud); Build production agents programmatically with SDK | | ||
| | **Model support** | Model-agnostic (Anthropic, OpenAI, and 100s others) | Claude models (Anthropic, Azure, Vertex AI, AWS Bedrock) | OpenAI models (GPT-5.3-Codex and variants) | | ||
| | **License** | MIT | MIT (underlying Claude Code is proprietary) | Apache-2.0 | |
There was a problem hiding this comment.
I am not sure if this is worth including, folks using deepagent will end up using a proprietary model as well
| | Skills System | ✅ [Skills](/oss/deepagents/skills) | ✅ [Skills](https://opencode.ai/docs/skills/) | ✅ [Skills](https://platform.claude.com/docs/en/agent-sdk/skills) | | ||
| | Long-term memory | ✅ [Memory Store](/oss/deepagents/long-term-memory) | ✅ [Rules](https://opencode.ai/docs/rules/) | ✅ [CLAUDE.md files](https://platform.claude.com/docs/en/agent-sdk/modifying-system-prompts#method-1-claude-md-files-project-level-instructions) | | ||
| | Streaming | ✅ [Streaming](/oss/deepagents/streaming/overview) | ✅ | ✅ [Streaming](https://platform.claude.com/docs/en/agent-sdk/streaming-output) | | ||
| | File Read/Write/Edit | ✅ [`ls`, `read_file`, `write_file`, `edit_file`](/oss/deepagents/harness#virtual-filesystem-access) | ✅ [Read, write, edit](https://platform.claude.com/docs/en/agent-sdk/overview#capabilities) | ✅ Built-in | |
There was a problem hiding this comment.
look at is here: https://langchain-5e9cc07a-preview-naomid-1771871114-3f4b6bb.mintlify.app/oss/python/deepagents/comparison#feature-comparison
I can turn it all into HTML but then the pure markdown is less readable...
| | Virtual Filesystems | ✅ [Virtual filesystem](/oss/deepagents/harness#virtual-filesystem-access) with pluggable backends | ❌ | ❌ | | ||
| | Composable Middleware | ✅ | [Hooks](https://platform.claude.com/docs/en/agent-sdk/hooks) | ❌ | | ||
| | Virtual Filesystems | ✅ [Virtual filesystem](/oss/deepagents/harness#virtual-filesystem-access) with [pluggable backends](/oss/deepagents/backends) | ❌ | ❌ | | ||
| <tr><td colspan="4"><strong>Agent Server</strong></td></tr> |
There was a problem hiding this comment.
In Agent Server, we could call this deployment potentially.
I'm not sure i'd classify codex app server as a deployment solution -- it looks a bit more like an adapter to some mcp like rpc? very hard to tell from the doc-page what it does, but it looks like it's meant to run over stdio, with experimental support for web-sockets. I don't really understand how one would do a standalone server that deploys a pre-configured agent
Protocols supported could be it's own category maybe?
For example ACP serer is separate from Agent Server -- it's not a deployment solution at all, but a way to hook up into an IDE primarily with an agent running on a local machine (at least at the moment)
MCP server mode -> this is more of a expose your agent as an MCP tool (at least for us).
✅ [LangGraph deployment](https://langchain-5e9cc07a-preview-naomid-1771871114-3f4b6bb.mintlify.app/oss/python/langgraph/deploy), self-hosted, or standalone
Do we call it Agent Server? LangSmith Deployment or LangGraph Deployment?
There was a problem hiding this comment.
I think LangSmith deployment
|
Mintlify preview ID generated: preview-naomid-1771968232-3e3dca0 |
| | Virtual Filesystems | ✅ [Virtual filesystem](/oss/deepagents/harness#virtual-filesystem-access) with pluggable backends | ❌ | ❌ | | ||
| | Composable Middleware | ✅ | [Hooks](https://platform.claude.com/docs/en/agent-sdk/hooks) | ❌ | | ||
| | Virtual Filesystems | ✅ [Virtual filesystem](/oss/deepagents/harness#virtual-filesystem-access) with [pluggable backends](/oss/deepagents/backends) | ❌ | ❌ | | ||
| <tr><td colspan="4"><strong>Deployment</strong></td></tr> |
There was a problem hiding this comment.
William FH (@hinthornw) would you mind helping us look at this comparison of production hosting solutions for the three products and provide a bit of feedback about feature gap?
Would you classify codex app server as a deployment solution?
There was a problem hiding this comment.
No I think codex app server is more a way to embed codex into your app vs. being able to deploy and customize codex as another app.
| **LangChain Deep Agents**: | ||
| - **Model flexibility**: Use any provider (Anthropic, OpenAI, or 100+ others) and swap models without changing your agent logic | ||
| - **Long-term memory**: Persist context across sessions and threads with the [Memory Store](/oss/deepagents/long-term-memory) | ||
| - **Sandbox-as-tool pattern**: Run individual operations in isolated sandboxes while the agent runs outside, or run the full agent inside a sandbox |
There was a problem hiding this comment.
Maybe worth mentioning sandbox provider flexibility here too. IMHO it's amazing you can swap sandbox provider easily
| | Agent runs in sandbox | ✅ [Can run in sandbox](/oss/python/deepagents/sandboxes#agent-in-sandbox-pattern) | ✅ [Can run in sandbox](https://opencode.ai/docs/#install) | ✅ [Can run in sandbox](https://platform.claude.com/docs/en/agent-sdk/hosting) | | ||
| | Agent runs in sandbox | ✅ [Can run in sandbox](/oss/python/deepagents/sandboxes#agent-in-sandbox-pattern) | ✅ [Can run in sandbox](https://platform.claude.com/docs/en/agent-sdk/hosting) | ✅ [Can run in sandbox](https://developers.openai.com/codex/security#os-level-sandbox) | | ||
| | Agent runs operations in sandboxes | ✅ [Uses sandboxes as tools](/oss/deepagents/sandboxes#sandbox-as-tool-pattern) | ❌ | ❌ | | ||
| | OS-level sandbox modes | ❌ | ❌ | ✅ `read-only`, `workspace-write`, and `danger-full-access` | |
There was a problem hiding this comment.
deepagent localshell sandbox allows you to let your agent go crazy on your own OS as well 🙈
| <tr><td colspan="4"><strong>Deployment</strong></td></tr> | ||
| | Production hosting | ✅ [LangGraph deployment](/oss/langgraph/deploy), self-hosted, or standalone | Build your own HTTP/WebSocket layer; run SDK in [containers](https://platform.anthropic.com/docs/en/agent-sdk/hosting) | [Codex App server](https://developers.openai.com/codex/app-server/) | | ||
| <tr><td colspan="4"><strong>Supported protocols</strong></td></tr> | ||
| | ACP server (IDE integration) | ✅ [ACP server](/oss/deepagents/acp) | ✅ (third-party) | ✅ (third-party) | |
There was a problem hiding this comment.
currently only support in Python but planning to ship for JS as well
There was a problem hiding this comment.
This got published: https://www.npmjs.com/package/deepagents-acp
| | Virtual Filesystems | ✅ [Virtual filesystem](/oss/deepagents/harness#virtual-filesystem-access) with pluggable backends | ❌ | ❌ | | ||
| | Composable Middleware | ✅ | [Hooks](https://platform.claude.com/docs/en/agent-sdk/hooks) | ❌ | | ||
| | Virtual Filesystems | ✅ [Virtual filesystem](/oss/deepagents/harness#virtual-filesystem-access) with [pluggable backends](/oss/deepagents/backends) | ❌ | ❌ | | ||
| <tr><td colspan="4"><strong>Deployment</strong></td></tr> |
There was a problem hiding this comment.
No I think codex app server is more a way to embed codex into your app vs. being able to deploy and customize codex as another app.
|
Mintlify preview ID generated: preview-naomid-1772032852-27e4dba |
|
Mintlify preview ID generated: preview-naomid-1772059447-fc3e358 |
|
Mintlify preview ID generated: preview-naomid-1772480186-e4594cd |
e3853ee to
7587f03
Compare
7587f03 to
e619d82
Compare
|
Mintlify preview ID generated: preview-naomid-1772629853-da7d320 |
| | **Use cases** | Custom general-purpose agents (including coding) | Custom AI coding agents | Prebuilt coding agent that can execute coding tasks | | ||
| | **Model support** | Flexible and model-agnostic (Anthropic, OpenAI, and 100s others) | Tightly integrated with Claude models (Anthropic, Azure, Vertex AI, AWS Bedrock) | Tightly integrated with OpenAI models (GPT-5.3-Codex and variants) | | ||
| | **Architecture** | Python SDK, TypeScript SDK, and CLI | Python SDK, TypeScript SDK | TypeScript SDK, CLI, desktop app, IDE extension, cloud interface | | ||
| | **Deployment** | LangGraph Platform | Self-hosted | Prebuilt agent that can execute tasks | |
There was a problem hiding this comment.
For OpenAI / Prebuilt agent that can execute tasks, can we clarify the deployment model here? Is this their cloud option / Codex web?
There was a problem hiding this comment.
I'm trying to express that you don't really deploy it. You install an app and then can get it to do tasks or you set up Codex web and then can use it in a repo. But you don't actually build and host agents. So maybe N/A is more applicable?
| | Additional instructions | ✅ [Memory](https://docs.langchain.com/oss/python/deepagents/customization#memory) | ✅ [CLAUDE.md files](https://platform.claude.com/docs/en/agent-sdk/modifying-system-prompts#method-1-claude-md-files-project-level-instructions) | ✅ [AGENTS.md](https://developers.openai.com/codex/guides/agents-md) | | ||
| | Long-term memory | ✅ [Memory Store](/oss/deepagents/long-term-memory) | ❌ | ❌ | | ||
| | Streaming | ✅ [Streaming](/oss/deepagents/streaming/overview) | ✅ [Streaming](https://platform.claude.com/docs/en/agent-sdk/streaming-output) | ✅ [Streaming](https://github.com/openai/codex/blob/main/sdk/typescript/README.md#streaming-responses) | | ||
| <tr><td colspan="4"><strong>Sandbox Integrations</strong></td></tr> |
There was a problem hiding this comment.
Should we highlight support for third party integrations (Daytona, Modal, Runloop)?
There was a problem hiding this comment.
Up to you and Harrison Chase (@hwchase17) , but I think it makes this more wordy, we'd need to list all available ones for each of them then. Claude Agent has 5 or 6 and then we need to keep them updated. Whereas linking to them is a bit easier
There was a problem hiding this comment.
If they're listed somewhere in each docs site could just link to them, but if it's not differentiating then not as important to add (I didn't know claude-agent-sdk supported those)
| | Skills System | ✅ [Skills](/oss/deepagents/skills) | ✅ [Skills](https://opencode.ai/docs/skills/) | ✅ [Skills](https://platform.claude.com/docs/en/agent-sdk/skills) | | ||
| | Long-term memory | ✅ [Memory Store](/oss/deepagents/long-term-memory) | ✅ [Rules](https://opencode.ai/docs/rules/) | ✅ [CLAUDE.md files](https://platform.claude.com/docs/en/agent-sdk/modifying-system-prompts#method-1-claude-md-files-project-level-instructions) | | ||
| | Streaming | ✅ [Streaming](/oss/deepagents/streaming/overview) | ✅ | ✅ [Streaming](https://platform.claude.com/docs/en/agent-sdk/streaming-output) | | ||
| | File Read/Write/Edit | ✅ [`ls`, `read_file`, `write_file`, `edit_file`](/oss/deepagents/harness#virtual-filesystem-access) | ✅ [Read, write, edit](https://platform.claude.com/docs/en/agent-sdk/overview#capabilities) | ✅ Built-in | |
There was a problem hiding this comment.
Should this just be read/write/edit for all 3?
| | Composable Middleware | ✅ | [Plugins](https://opencode.ai/docs/plugins/) | [Hooks](https://platform.claude.com/docs/en/agent-sdk/hooks) | | ||
| | Pluggable Storage Backends | ✅ [backends](http://localhost:3000/oss/python/deepagents/backends) | ❌ | ❌ | | ||
| | Virtual Filesystems | ✅ [Virtual filesystem](/oss/deepagents/harness#virtual-filesystem-access) with pluggable backends | ❌ | ❌ | | ||
| | Composable Middleware | ✅ | [Hooks](https://platform.claude.com/docs/en/agent-sdk/hooks) | ❌ | |
There was a problem hiding this comment.
Do we want ✅ for Hooks?
|
Mintlify preview ID generated: preview-naomid-1772645021-600176c |
|
Mintlify preview ID generated: preview-naomid-1772645466-88e84d5 |
|
Mintlify preview ID generated: preview-naomid-1772715388-2ed6d7d |
e9ca004 to
9482a08
Compare
|
Mintlify preview ID generated: preview-naomid-1772797977-749ec98 |

https://langchain-5e9cc07a-preview-naomid-1772715388-2ed6d7d.mintlify.app/oss/python/deepagents/comparison