Skip to content

Agent-Reach Review

Agent Reach is a free, MIT-licensed CLI that lets AI coding agents such as Claude Code and Codex search and read YouTube, Reddit, X, GitHub, Bilibili, XiaoHongShu and more.

Free · API

Stork's verdict on Agent-Reach

Agent Reach is the easiest free way to give a CLI agent read access to social and Chinese platforms, as long as you treat it as a research helper, not a pipeline.

Agent-Reach reviewed by Stork AI · stork.ai/en/agent-reach

stork
I am one still image.
One typed sentence. This clip cost 5 cents, the minimum per video.
Make a photo talk for 20 cents a minute.
Upload one photo and type the words. No subscription. One free video a day.
Make one freeDevelopers: there is an API

What Agent Reach does

Agent Reach is an open-source command-line tool, published on GitHub under the MIT license, that gives CLI agents such as Claude Code, Codex, Hermes Agent and OpenClaw access to platforms their built-in web search usually cannot reach: pages behind a login (XiaoHongShu notes, an X account's latest posts, job listings), subtitles and speech-to-text for YouTube and Bilibili, Chinese platforms like V2EX and Xueqiu, and structured GitHub repos, issues and pull requests. Your agent calls it as agent-reach <channel> <action>, for example agent-reach youtube search "LLM agents", or agent-reach read <url> to pick the channel from a link. Each action has an ordered list of upstream tools (such as yt-dlp, bili-cli, OpenCLI, Jina Reader and Exa); if the first fails, it falls back to the next.

Setup is done by the agent itself: you paste the project's install.md link into your agent, and it installs uv if needed, the program and the default channels. Everything lives in ~/.agent-reach/, and you manage it in plain language ("add Xiaohongshu", "upgrade Agent Reach", "uninstall Agent Reach"). For login channels it runs its own background browser and copies that one site's login from your Chrome; the project says it never reads your password store and never logs in for you, and it recommends a secondary account. It requires Python 3.10 or later. The project warns that the PyPI package named agent-reach is not this project; install only from the GitHub URL in install.md. The current release is v1.6.1 (November 1, 2026).

Agent Reach pricing

PlanMonthly priceWhat you get
Open source (MIT)$0The full CLI, every channel, self-installed on your machine

Agent Reach has no paid plan, hosted service or account. The project describes itself as "zero API fees": the default channels (web, YouTube, GitHub, V2EX, Bilibili) need no keys. Two things can still cost money outside the project: YouTube speech-to-text needs a Groq or OpenAI API key, and Xiaoyuzhou podcast transcription needs a Groq key, both billed by those providers at their own rates. Login channels use your own accounts on each platform. The README lists paid sponsors (proxy and scraping vendors); they are optional and not required to run the tool.

Prices checked on the vendor's pricing page on November 10, 2026. Agent Reach has no pricing page; terms are in the GitHub repository and README.

Key features

What the project's README lists for the current version:

  • Default channels that work right after install: web, YouTube, GitHub, V2EX, Bilibili
  • Optional channels: XiaoHongShu, Xueqiu, Boss Zhipin, Xiaoyuzhou, LinkedIn, Reddit
  • Experimental channels: X (Twitter), Facebook, Instagram
  • YouTube and Bilibili subtitles, plus speech-to-text with a Groq or OpenAI key
  • Ordered fallback tools per action, so a broken upstream tool can be swapped without changing commands
  • Status check that shows which channels work now and which broke since last time
  • Read-only by design, with a dedicated background browser reachable only from your machine
  • Plain-language install, upgrade and uninstall handled by your agent

Who Agent Reach is best for

Agent Reach suits developers and researchers who already work in a CLI agent and want it to read social platforms, video transcripts and Chinese sites without wiring up separate APIs. It is especially useful for Chinese-language sources (Bilibili, XiaoHongShu, V2EX, Xueqiu), which general web search covers poorly. It is a local, read-only research helper, not a data pipeline: login channels depend on your own accounts and on scrapers that platforms can block at any time, and the X, Facebook and Instagram channels are marked experimental.

If you need a hosted, API-key-based crawler for production, look at Firecrawl or Apify. For an agent-friendly search API, Tavily is the usual pick, and for a self-hosted open-source crawler, Crawl4AI.

What users say

Third-party coverage is mostly write-ups rather than long-term use reports. Wavect's open-source review (written by an AI engineering consultancy) calls it "an excellent low-cost prototype layer" and credits it for showing channel health, but says the zero-cost claim is "incomplete" because some channels need cookies, keys or proxies, and that it is "not a drop-in production data platform". The same review warns that account-backed use risks platform restrictions and that teams must check each platform's terms before scraping. AI IDE List's review praises its mix of Western and Chinese platforms and its fit with terminal agents, but warns that cookie-based logins expire and that commands can succeed while returning empty data. Reviewers agree it is a strong starting point for agent research workflows, with reliability and legal review left to the user.

Pros

  • Free and MIT-licensed, with no account or hosted service
  • Covers Chinese platforms like Bilibili, XiaoHongShu and V2EX that search tools miss
  • Read-only by design, with a local-only dedicated browser
  • Fallback tools per action keep commands stable when an upstream tool breaks
  • Agent handles install, upgrade and clean uninstall

Cons

  • X, Facebook and Instagram channels are experimental and may stop working
  • Login channels use your own accounts, which platforms may restrict
  • Speech-to-text needs a paid Groq or OpenAI API key
  • Beta-stage project, not a production data platform

Frequently asked questions

Is Agent Reach free?

Yes. Agent Reach is open source under the MIT license and has no paid plan. The default channels need no API keys. YouTube speech-to-text needs a Groq or OpenAI key and Xiaoyuzhou transcription a Groq key, billed by those providers.

Which platforms does Agent Reach support?

Default: web, YouTube, GitHub, V2EX and Bilibili. Optional: XiaoHongShu, Xueqiu, Boss Zhipin, Xiaoyuzhou, LinkedIn and Reddit. Experimental: X (Twitter), Facebook and Instagram, which the project says may break at any time because of stricter anti-scraping.

How do I install Agent Reach?

Send your agent the line from the README: install Agent Reach from https://raw.githubusercontent.com/Panniantong/Agent-Reach/main/docs/install.md. The agent installs uv if missing, then the program and default channels. The same command upgrades it. Do not install the PyPI package named agent-reach; the project says it is not theirs.

Which AI agents work with Agent Reach?

The README says any agent that can run commands can use it, and names Claude Code, Codex, Hermes Agent and OpenClaw. If your agent says it cannot run commands, enable command execution in its settings.

Does Agent Reach access my accounts or passwords?

For login channels it uses its own background browser and copies that one site's login from your Chrome; the project says it does not change your browser, does not read your password store and never logs in for you. It recommends switching to a secondary account, since platforms may restrict accounts used for automated access.

Can Agent Reach post on social media?

No. The project describes it as read-only: it searches and reads, and never posts, comments, likes or sends messages.

Compare Alternatives

See every Agent-Reach alternative, compared

Agent-Reach

Free · API

From the Stork blog

Enjoying this? Get one like it in your inbox each morning.

one email a day · unsubscribe in two clicks · no third-party tracking