Agent comparisons
Gemini Spark vs Hermes Agent: Workspace or Your Runtime
Choose Gemini Spark when the work sits in Gmail, Calendar, Drive, and Docs and you want Google to run that agent. Choose Hermes Agent when you want to pick the model and keep the task instructions in files you can read. A managed Hermes instance stays online after you close your laptop. It does not arrive connected to those Google apps, and a Google model key does not change that.
Gemini Spark is the name used in this comparison. On October 20, 2026, its former product URL redirects to Google’s general Gemini page; current help describes tasks in Gemini. That wording change does not establish a formal rebrand. Hermes Agent is Nous Research’s open-source runtime, distinct from the Hermes language-model family.
How do Gemini Spark and Hermes Agent compare?
| Decision | Gemini Spark | Managed Hermes |
|---|---|---|
| What you start with | An agent inside the Gemini app | An instance you configure |
| Connected Google apps | Connected apps, subject to account permissions | No native Workspace connectors |
| Unattended time | Schedules: time-based or conditional triggers | No managed cron on Hermes |
| Laptop closed | Remote browsing can continue; local Chrome needs an awake device | The hosted instance keeps running |
| Browser | Chrome auto browse can use logged-in accounts and saved passwords, with permission | Optional Hosted Browser; you sign into a shared profile |
| Model | Gemini model choices in the app | A provider you connect, including Google's API if you choose it |
| Before a risky action | Documented browser confirmations and human takeover | Messaging allowlists are not a per-action approval |
| Price | Google AI Pro or Ultra; eligibility restrictions apply | Hosting from $6.90/month; model access separate |
The Google column follows the help pages linked in each section. The Hermes column is Lobsterland’s managed Hermes hosting, not every option in the upstream project.
When is Gemini Spark the better fit?
Choose Google when app access is the main reason you want an assistant. A useful evaluation is an inbox-to-calendar task using accounts you can actually authorize. Compare the resulting permissions and output, not just whether both products can summarize text.
Hermes is the stronger candidate when you want to configure the model, inspect reusable instructions, and talk to the runtime through your chosen channel. Hosting removes server administration; connecting Google as a model provider does not configure Gmail. Browser access or a separate integration needs its own authorization.
The decision is therefore about the workflow you are buying. A model that writes a good summary is only one part of an assistant that can find the source email, update a file, and deliver the result.
Does managed Hermes keep running when your laptop is off?
Yes. The hosted Hermes instance keeps running when your laptop closes. Its tools still need the right credentials. A running process does not acquire an inbox or a website session just because it is online.
A Hermes instance can answer in built-in chat, Telegram, or Slack. Its optional Hosted Browser has a separate profile. Signing into Gmail there can expose that website to the agent; signing into Gmail on your laptop does not. Neither path should be confused with a native Workspace connector included in hosting.
How does browser access differ?
Google’s auto browse documentation distinguishes two browser paths. Local Chrome can use your existing sessions and permissioned saved sign-ins, but needs the device awake. Remote browsing can continue after you close it and can pause for you to sign in. Auto browse has separate device, region, language and plan requirements; task eligibility alone does not establish access.
Managed Hermes does not attach to your laptop’s Chrome profile. The optional Hosted Browser runs beside the instance. You sign into the site there, and the agent uses that same persistent profile through its configured browser connection. Laptop cookies and saved passwords are not imported. Decide which accounts belong in that profile before asking the agent to act through it.
Google documents plan review, confirmation and takeover for browser tasks. Those are product controls, not proof that every action is safe. We have not tested them in an account. Managed Hermes does not reproduce that Google interface or its confirmation policy.
Does a Google model key replace Gemini Spark?
No. Lobsterland can connect Hermes to Google as a model provider, alongside Anthropic, OpenAI, OpenRouter, and MiniMax. The credential configures inference: prompts go to the model API. It does not authorize access to your Gmail, Calendar, or Drive.
Using a Gemini model inside Hermes is closer to OpenClaw's Gemini provider setup than to Spark: you are choosing an inference vendor. You are not granting the agent your Drive. If the task needs a specific Google file, paste the excerpt or export it. Do not assume the model name "Gemini" unlocked the apps.
Treat model credentials, app authorization, and browser sign-in as three separate decisions. If you need only one document summarized, sharing a redacted excerpt may be sufficient. If you need ongoing app access, verify the integration and permissions before choosing the runtime.
What do schedules, skills, and confirmations cover?
Google’s schedule help documents time-based jobs, Gmail-filter triggers, and topic monitors. It cautions against relying on monitors for time-critical events. Its skills help describes reusable task instructions. A saved instruction and a trigger that runs it are separate capabilities.
Hermes skills on Lobsterland are inspectable SKILL.md files persisted with the instance. The Hermes skills guide covers installation. Installing instructions grants no Gmail access and creates no schedule. Managed Hermes has no cron surface; upstream automation options are a different scope. If scheduling is required, compare the hosted runtimes in OpenClaw vs Hermes. OpenClaw’s scheduler still does not supply Google app authorization.
Lobsterland configures Hermes tool approvals with mode: off. Telegram and Slack allowlists control who may message the agent; they do not require review of each tool action. A prompt asking the agent to stop before sending is an instruction, not an enforced transaction boundary. Evaluate that distinction before giving a workflow an account that can send or modify data.
What do access and cost look like?
As checked on October 20, 2026, Google’s task eligibility requires a personal account, age 18+, Pro or Ultra, Keep Activity, and Personal Intelligence. Work and school accounts are excluded. So are the EEA, Nigeria, Switzerland, and UK. Access to Workspace apps does not mean eligibility with a Workspace account. Check the requirements for your account before paying.
Lobsterland's Hermes hosting starts at $6.90 per month. Every plan includes a seven-day trial: add a card up front, and the hosting charge happens when the trial ends unless you cancel first. Model access and optional add-ons are separate. The hosting pricing page lists the capacity tiers. A $6.90 instance plus a Google API bill is not a substitute price for a Google AI plan, because it does not include Spark.
How should you test Hermes against a Workspace-shaped task?
Use a task that does not pretend the agent has your inbox. Paste a redacted thread you are willing to share, then ask: "List the decisions and the open questions. Quote the line each one came from. Draft a reply I can edit. Do not send anything." This is an evaluation prompt, not a measured result.
- Launch Hermes and connect one model. Run the prompt in built-in chat. If you pick Google here, you are only choosing the model API.
- Check the quotes. Every decision should point at a line you pasted. A confident summary with no quote is a failed check, not a finished assistant.
- Save the instructions. When the shape is right, turn it into a skill and read the file. The next run should follow that file, not a hidden connector.
- Add one extra surface only if the task needs it. Telegram or Slack changes where you talk to it. Hosted Browser changes which website session it can see. Neither grants Gmail access automatically; a signed-in browser can expose the site to the agent.
Keep the rest of the surface honest. Managed Hermes includes skills, logs, upgrades, Paperclip, and optional Hosted Browser. The native dashboard is available from image v2026.6.5. It does not include managed cron, a workspace-file editor, multi-agent routing, usage analytics, or WhatsApp. Spark wins when those Google apps and schedules are the product you want. Hermes wins when you would rather configure the model and the method than hand Google the apps.
People comparing other personal agents can read Instinct vs Hermes Agent. That product is a different assistant, not a Workspace connector.
Frequently asked questions
Is Gemini Spark the same thing as Gemini?
This comparison uses the Spark name for Google's agent workflow. Google's current help calls the workflow tasks. A Gemini model API is a separate way to use a model; it does not include that app workflow.
Can a Google API key make Hermes read Gmail?
No. On Lobsterland, Google is one supported model provider. That connection sends prompts to Google's model API. It does not turn on Gmail, Calendar, Drive, Docs, Sheets, Slides, YouTube, or Maps.
Does managed Hermes keep working when my laptop is off?
The hosted instance keeps running after you close your laptop. That does not give it Gemini Spark's Workspace connectors or Chrome saved-password browsing. A task still needs the tools you actually configured.
Does Gemini Spark require a paid plan?
Yes. Google's task help checked October 20, 2026 requires Google AI Pro or Ultra. Other account, age, settings and regional requirements apply; see the access section above.