How to connect Gong 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 Gong from the dashboard, then start a fresh chat if the runtime still has the old tool catalog loaded.

Gong 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 Gong 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.

Gong 403s are about what the credential can reach, and there are two layers. The layer search results teach applies to API tokens: a token generated with limited data scopes fails endpoints outside those scopes, so the fix is regenerating the token with the right scopes. The layer that applies to a hosted OAuth connection is Gong's own permission model: the connected user's role and workspace decide what the agent can see, so a call or folder outside that user's visibility returns a permission error rather than a not-found. On top of both sits the page's own cause, a placeholder ID such as YOUR_CALL_ID reaching the API literally. Ask the agent which user it is acting as and which object it was trying to reach before anything else.

Ask the agent to diagnose it:

Call gong_get_user_by_id and gong_list_all_company_workspaces_v2_workspaces and tell me which user and workspaces this connection can see. Then quote the exact error from the failed call. Do not retry it yet.

Gong's documented API limits are 3 requests per second and 10,000 requests per day, with a 429 and a Retry-After header when exceeded, and the allowance is company-level rather than per connection. A 57-tool agent walking folders and transcripts one call at a time is exactly the shape that can plausibly reach the daily pool. The response is to change the shape of the work: gong_retrieve_call_data_by_date_range_v2_calls pulls many calls in one request instead of one call at a time, name the folder or date range up front, and have the agent read and report the Retry-After value instead of retrying blind. These limits are Gong-side and documented on their help center; no reconnect changes them.

Ask the agent to diagnose it:

Tell me how many Gong calls you have made in the last minute and what you were enumerating. Quote the exact error text and any Retry-After header. Do not call Gong again until I reply.

Two failures share this symptom. If no Gong tools appear at all, the connection or pairing is incomplete: confirm the plugin was installed with --enable, that hermes clawlink begin and hermes clawlink finish both ran, and that Gong shows as connected in the dashboard. If most tools work and one name fails, the name is wrong rather than missing, and the error lists the closest real ones. There is also a first-call timing case: schemas load on demand, so the opening Gong call in a fresh session can arrive before the catalog and needs one retry. The explanation search results give for this symptom, a client that did not load the MCP config, belongs to a local server setup and does not apply here.

Ask the agent to diagnose it:

List the Gong tools you actually have access to. If there are none, say so plainly and tell me whether Gong shows as connected. Do not guess tool names.

Try reconnecting Gong and complete the consent flow in the same browser session. Partial OAuth approvals or switching accounts mid-flow can leave the connection incomplete.

Yes. ClawLink is the fastest way to connect Hermes to Gong: link your Gong account once in the browser and Hermes Agent can call the Gong API through 57 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 Gong in the dashboard and Hermes can use it from the next message — no config files or API keys to manage.

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

The alternative to ClawLink is usually manual OAuth app 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 Gong working from chat.

Gong's official MCP server is hosted, read-only, and created and authorized by a Gong administrator, and it answers questions about accounts and deals through AI-generated summaries rather than exposing raw data to the client. This page's route is a different shape: your own Gong login authorizes the connection, the agent gets 57 named tools covering calls, transcripts, users, workspaces, scorecards, Engage flows, and CRM objects, and raw transcripts are retrievable through gong_get_call_transcript. The official server is the right choice if you want Gong's own summaries inside a plain MCP client. This route is the right choice if you want Hermes to work with the underlying data directly, without an administrator creating the integration.

No, and that is the point of the hosted OAuth route. The manual path documented on Gong's help center requires Technical Administrator privileges in the workspace to generate an Access Key and Secret, which puts the API behind exactly the people who least want to set up agent connections. Here you connect with your own Gong login and the agent's access follows your own permissions in the workspace. If your account can see a call, the agent can read it; if it cannot, a permission error is Gong enforcing that, not a broken connection.

Yes, through gong_get_call_transcript and the call data tools, and this is the capability to be deliberate about. Every safety answer in the search results flags transcripts as sensitive customer and employee data that deserves least-privilege handling. The hosted connection respects your Gong account's own visibility, so the agent cannot see more than you can, but that is exactly why prompts should name the call or folder they need rather than asking for everything. Transcripts are the class of data where a wide prompt is a real privacy decision, not a performance question.

The call and revenue surface: call metadata and transcripts, calls by date range and by folder, interaction statistics, user activity statistics, users and workspaces, permission profiles, scorecards and coaching metrics, Engage flows and flow folders, tasks, call outcomes, CRM objects and integration details, and data privacy information for phone numbers and email addresses. It is read-heavy by design, and the writes that exist, like gong_manage_user_call_access, change access for real people, so they deserve a confirmation step.

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.