How to connect Apify to Hermes Agent

Pairing is a two-step handshake: run hermes clawlink begin, approve the link in your browser, then run hermes clawlink finish. If you ran finish before approving, or the approval link expired, run hermes clawlink begin again to get a fresh link. Confirm the plugin was installed with --enable, then verify with hermes clawlink test.

Reconnect Apify from the dashboard, then start a fresh chat if the runtime still has the old tool catalog loaded.

Apify tool schemas load on demand the first time a tool runs and are cached after that, so this error usually clears on its own: wait a few seconds and retry the same request. If every Apify call keeps failing with it in a fresh chat, reconnect from the dashboard, and contact support if it still persists — that pattern points to a configuration problem on our side, not something you can fix by reconnecting again.

Two usual causes. The connected account may not have access to the specific workspace, inbox, store, or project in the request — check that first. If access looks right, the agent may have sent a placeholder value (like "YOUR_ID" or an example id from documentation) instead of a real one: ask it to run a list or search tool first, then retry the action with a real id from those results. Most failures at this stage are one of these two, not ClawLink bugs.

Because this is a token integration, start with the token itself rather than with account access. Apify tokens carry scopes, and a scoped token that can read datasets may be unable to start Actor runs, which produces a connection that looks healthy and fails only on writes — the answer is a token with the permissions the work needs, re-pasted from the dashboard. A rotated or deleted token in Apify produces a hard failure across every call instead, which is the easiest case to recognise. Once the token is ruled out, the remaining cause is resource ownership: Actors, tasks, and datasets belong to accounts, so a run or dataset id from someone else's account returns a permission error rather than a not-found, as does a placeholder id copied from documentation.

Ask the agent to diagnose it:

Call apify_users_me_get and tell me which Apify account this connection is acting as. Then run apify_get_list_of_tasks and tell me whether it succeeds. Do not retry the failed call yet.

Two different problems that look alike. If Hermes shows no Apify tools whatsoever, pairing or the connection is the issue: confirm the plugin is installed with --enable, that hermes clawlink begin and hermes clawlink finish both completed, and that Apify shows as connected in the dashboard. If some tools work and one name fails, that name is wrong rather than missing, and the error lists the closest real ones. There is also a timing case unique to this setup: tool schemas load on demand rather than shipping with the plugin, so the very first Apify call in a fresh session can arrive before the catalog does — retry once. None of this involves editing a client config file, which is what most search results for this symptom describe.

Ask the agent to diagnose it:

List the Apify tools you actually have access to. If there are none, tell me that plainly. If there are, tell me which one starts an Actor run and use that exact name.

Almost always the run-versus-dataset distinction. Starting an Actor and reading its output are separate calls, so an agent that runs apify_run_actor and stops has a successful run and no records. Ask it to take the run's dataset id and call apify_get_dataset_items. The second cause is timing: apify_run_actor returns as soon as the run starts, so reading the dataset immediately gives you an empty or partial set for anything that takes more than a moment. Either use apify_run_actor_sync, which waits, or poll the run's status before reading. If the dataset really is empty, apify_get_log for that run tells you what the Actor did.

Ask the agent to diagnose it:

Give me the run id and its status, then call apify_get_dataset_items for that run's dataset and tell me how many items came back. If it is zero, pull apify_get_log for the run.

Double-check that the API key for Apify has the right scopes or account access. A valid key can still be too limited for some reads or writes.

Yes. ClawLink is the fastest way to connect Hermes to Apify: link your Apify account once in the browser and Hermes Agent can call the Apify API through 112 ready-made tools — no custom code or token handling.

Install the plugin with hermes plugins install ClawLink-HQ/hermes-plugin --enable, then pair once: run hermes clawlink begin, approve the link in your browser, and run hermes clawlink finish. Connect Apify in the dashboard and Hermes can use it from the next message — no config files, and the Apify key you paste is stored server-side instead of in your environment.

About two minutes. Sign in, click Connect next to Apify in the dashboard, authenticate, and Hermes Agent can use it from the next chat message.

The alternative to ClawLink is usually manual API key setup plus your own token handling, permission troubleshooting, and tool plumbing for Hermes Agent. That is fine if you want to build and maintain the integration yourself. Most teams just want Apify working from chat.

You paste a key, once, on the hosted setup page — Apify authenticates with an API token rather than OAuth, so anything promising a keyless Apify connection is describing something else. What changes is where the token ends up. Instead of sitting in an environment file on your machine, or in a config for an MCP server you run, it is stored server-side and used on your behalf when the agent calls a tool. You can revoke it from the ClawLink dashboard, and rotating it in Apify means updating it in one place rather than everywhere you deployed it.

Run them. There are separate tools for the two shapes that matter: googling this usually turns up only the async pattern, but here apify_run_actor starts a run and returns immediately, while apify_run_actor_sync waits for the run to finish and hands back the result. The synchronous one is what you want in chat when the Actor is quick, because the agent can act on the output in the same turn. For long scrapes, start the run, then poll with apify_get_list_of_runs and pull results from the dataset when it completes.

Through the dataset, not through the run. An Actor run produces a dataset, and apify_get_dataset_items is what reads the rows out of it — a common first mistake is expecting the run tool itself to return the scraped records. For Actors that write to a key-value store instead, apify_get_key_value_record fetches a named record. If a run looks successful but the agent reports no data, the dataset is the place to look before assuming the Actor failed.

It can, and this is the thing worth being deliberate about, because Actor runs consume compute units on your Apify account exactly as they would if you started them yourself. An agent asked an open-ended question may start several runs while exploring. Two habits help: name the Actor and the input explicitly rather than asking the agent to find something that scrapes X, and have it check apify_users_me_usage_monthly_get when you want to know where the month stands before launching anything large.

Pairing is a two-step handshake: run hermes clawlink begin, approve the link in your browser, then run hermes clawlink finish. If you ran finish before approving, or the approval link expired, run hermes clawlink begin again to get a fresh link. Confirm the plugin was installed with --enable, then verify with hermes clawlink test.