A one-command 'migrate from OpenClaw' tool; the thread can't find the use case
Hermes Agent added a single command to import your whole OpenClaw setup - persona, memory, skills, even your Telegram and Signal tokens. The top comment isn't about migration. It's a plagiarism link.
Nous Research's Hermes Agent shipped a guide and a 'hermes claw migrate' command to absorb OpenClaw configs, MEMORY.md, skills and messaging tokens in one shot. The first reply instead pointed at a similarity analysis accusing Hermes of copying a rival's memory/skill/evolution design.
Then dozens of users who actually ran these self-hosted agents asked the same thing: what is any of this for? One had built his own clone in a few hours; another said his version calls his wife if he ignores a pull request.
A migration path between two agents nobody can quite justify
Nous Research published a migration guide for moving from OpenClaw to its own Hermes Agent. The mechanics are unusually complete: a hermes claw migrate command previews and then imports the persona file (SOUL.md), workspace instructions (AGENTS.md), long-term memory (MEMORY.md, USER.md), skills, MCP servers, text-to-speech voices, and access tokens for Telegram, Discord, Slack, WhatsApp, Signal, Matrix and Mattermost - with dry-run, backup, and a hermes claw cleanup to archive the old install. OpenClaw is the predecessor self-hosted "agent that lives on your machine"; Hermes is the more polished successor.
The thread opened not on migration but on provenance.
The first reply is an accusation, not a how-to
The top comment steered straight into a dispute over whether Hermes copied a competitor.
"It's probably good to at least be aware of the plagiarism debacle around Hermes." — a commenter linking the Hermes plagiarism dispute
The submitter himself sharpened it, pointing to the specific claim: that Hermes shipped a "strikingly similar memory / skill / evolution-asset design" to a rival's Evolver without attribution.
"Linking to the claim might be stronger." — JumpCrisscross, the story's submitter, surfacing the evomap.ai similarity analysis
A skeptic deflated the drama: none of this is novel.
"there is no debacle… there is nothing new about memory (aka sessions saved), skills (aka prompt in a file), etc folks have been doing this for 2+ years." — a builder who lists prolog, lisp, APL and forth in his bio
The real fight: what is any of this actually for?
Migration aside, the thread's center of gravity was a blunt question - what do you do with these things? A deeply technical reader laid out his stack and still came up empty.
"I'd love some OpenClaw master to opine on meaningful use cases beyond clawbook, checking the weather, and telling you about crypto news, as it genuinely feels like something I should find utility for but am just too old or something." — a commenter running local 4090s, fine-tuned models and a custom TTS pipeline
Several had simply built their own. One described a harness with teeth:
"If it sees a problem in email it will create a ticket, assign it to an agent, monitor progress, get a PR ready for me - then it harasses me until I review the PR. If I ignore the PR it will call my wife." — a user who full-shot a prompt to clone the whole thing
The sharpest dismissal questioned the market itself, and named where the energy went:
"Grifters, FOMO driven technophiles, and wannabe 'founders' jumped on the bandwagon… Now that the creator got cushy job at OpenAI, and the hype is over, it's just another tech bro toy to burn tokens with" — a commenter on the size of the technophile market
The one concrete win came from outside the terminal
Amid the shrug, the clearest endorsement wasn't a developer toy at all. A product lead wired Hermes into company data so non-engineers could self-serve.
"I set up a Hermes agent for our company and connected it to BigQuery, Stripe, Amplitude, etc so that our nontechnical teammates can ask and answer their own data questions via Slack. It works great!" — Head of Product at Once, formerly Filepicker.io
And Simon Willison answered the "who actually makes money" question with the oldest truth in tooling: "Users who figure out ways to use their agents that are profitable to them make money. Everyone else spends money." That is the real shape of the story. A slick one-command migration path is the easy part; an industry of OpenClaw, NanoClaw, Hermes, Pi and a dozen forks has nailed the plumbing - persistent memory plus cron - while most of its loudest users still can't name a task that justifies the token bill. The strongest reviews here describe agents people built themselves in an afternoon, not the ones they migrated into.
Get breakdowns like this every day
The top Hacker News stories, unpacked — by email, Telegram, or RSS.
A '90% token savings' AI tool gets called an illusion. Its author confesses why
RTK compresses terminal output for AI coding agents and screenshots a big 'gain' number. A skeptic argues that gain measures the wrong thing entirely. Then the author of the takedown appears with the real reason he wrote it.