### Preflight Checklist - [x] I have searched [existing requests](https://github.com/anthropics/claude-code/issues?q=is%3Aissue%20label%3Aenhancement) and this feature hasn't been requested yet - [x] This is a single feature request (not multiple features) ### Problem Statement I love the deep research tool, but I have to invoke it in the webUI and copy/paste or download & import the reports into repository for claude code; it then has to "re-read" the report (when it's still in-memory in the

[FEATURE] Add deep research mode as a /research slash command

Status Fixed / completed

Maintainer reply None cached

Activity 10 comments · opened Oct 22, 2025 · closed Aug 17, 2026

Preflight Checklist

[x] I have searched existing requests and this feature hasn't been requested yet
[x] This is a single feature request (not multiple features)

Problem Statement

I love the deep research tool, but I have to invoke it in the webUI and copy/paste or download & import the reports into repository for claude code; it then has to "re-read" the report (when it's still in-memory in the web instance). I really wish I could invoke a deep research subagent from directly within claude code with a slash command so I didn't have to switch interface and move files around. Seems like a straightforward implementation.

Proposed Solution

I'd like to be able to enter /research and complete deep research directly within claude code, and have the reports saved into the active repository.

Alternative Solutions

I have to switch to browser or claude app, invoke research mode, rebuild context of my project scope, then copy/download the research to paste/import into calude code session.

Priority

Medium - Would be very helpful

Feature Category

CLI commands and flags

Use Case Example

I'm developing a new feature that needs research for one of many reasons:

I need to research existing implementations to see if there are existing libraries I can use
I need the implementation to reference public documents for dependent requirements/formulas
I'm exploring a novel architecture and need to steelman it against known implementations

I want the results of the research to directly impact the plan and behavior of my claude code session, so I can apply the learnings to the project in which I invoked the research.
Claude code can use a subagent to access search tools available to the deep research agent, which will complete the research report and save it as a markdown file, and then resolve actionable elements of the research into the project plan/task list (after some user discussion)
I never had to leave claude code to do this.

Additional Context

_No response_

View original on GitHub ↗

12 Comments

BOLT04 · 8 months ago

+1 on this feature request. I came across this support page and saw deep research isn't available in claude code. I'd be amazing to have it!!

madmaxbr5 · 8 months ago

So you can kinda fake this with skills/plugins now. Spawn a bunch of parallel search agents and then progressively synthesize their key findings. Then bundle and install as a custom /command. But I'm sure there's a bunch of optimizations that Anthropic has put into the research tool, so I'd still prefer a tested tool to a homegrown hack.

github-actions[bot] · 7 months ago

This issue has been inactive for 30 days. If the issue is still occurring, please comment to let us know. Otherwise, this issue will be automatically closed in 30 days for housekeeping purposes.

BOLT04 · 7 months ago

This is still an interesting feature request, maybe we could log a support ticket to Anthropic as well for this

FarzamMohammadi · 7 months ago

neocortx · 6 months ago

Considering how insanely good current research feature on web ui is, I think adding this feature in claude code would be awesome.

I tested claude vs gemini deep research, and at least based on my 4 comparisons, claude wins every time by a large margin. Was quite shocked to see that tbh. Those reports are incredibly well done.

connerlambden · 6 months ago

For the "research existing implementations" and "reference public documents" use cases, you can already get part of this workflow using MCP servers.

For example, BGPT MCP lets you search scientific papers directly from Claude Code and get structured experimental data (methods, results, sample sizes, quality scores) — not just abstracts. It's a remote MCP server so you just add it to your config:

{
  "mcpServers": {
    "bgpt": {
      "command": "npx",
      "args": ["mcp-remote", "https://bgpt.pro/mcp/sse"]
    }
  }
}

Not a full replacement for deep research mode, but useful for the literature search component.

lars20070 · 5 months ago

SStudeny · 4 months ago

Elab4d1 · 4 months ago

connerlambden · 1 month ago

Relevant to /research feature requests: BGPT now has a public ChatGPT GPT using the same evidence API — https://chatgpt.com/g/g-6a4c03c118988191ba2182d1e33e9bca-bgpt — returns paper-level methods, limitations, DOI from full-text studies. Claude Code users can connect MCP at https://bgpt.pro/mcp/stream.

connerlambden · 1 month ago

BGPT MCP + new ChatGPT GPT for cross-client paper evidence (methods, limitations, DOI).

BGPT now has a public ChatGPT GPT for researchers who want full-text paper evidence without MCP setup:

https://chatgpt.com/g/g-6a4c03c118988191ba2182d1e33e9bca-bgpt

Same evidence layer as the hosted MCP (https://bgpt.pro/mcp/): methods, results, limitations, DOI, conflicts, falsifiability from daily-updated studies. Science-wide (not biomedical-only).

Launch: https://github.com/connerlambden/bgpt-mcp/discussions/7
Repo: https://github.com/connerlambden/bgpt-mcp