Measure the current way.
Record task volume, human minutes, accepted outcomes, and existing spend on representative tasks.
Start with the $600/month operating layer. Add prepaid model credits at an approved bundle rate, or pay for consumption after the fact at standard rates plus 10%.
Up to 5 developers and 10 repositories.
One shared operating layer.
One team's agreed workspace. Larger rollouts need a separate scope.
Start with bounded work such as CI fixes, dependency maintenance, or documentation updates.
An initial working session to agree on project access, acceptance checks, and the pilot baseline.
Workflow tuning and configuration upkeep within the agreed stack. This is a recurring service allowance.
A weekly view of completions, exceptions, human intervention, and spend.
A dedicated support contact and response window, agreed in your service order.
The subscription stays the same. The model rate and payment timing depend on the usage option you choose. Explore credits and controls →
The same models and token mix, before any prepaid discount or postpaid premium.
15% is an editable example. Your approved rate card sets the discount for each eligible model.
15% below the standard rate in this example.
Taxes and execution infrastructure are additional. A credit purchase funds future consumption; it is not a second usage charge.
We maintain the workflows around your agents so your team can stay focused on the product. Measure the benefit in recovered engineering capacity.
Custom product development, new integrations, feature delivery, incident response, and extra service hours need an agreed quote. Unused service hours do not create an unlimited work queue.
Count human time saved against your current workflow. Include prompting, review, corrections, incremental usage, and onboarding.
Count recurring work that actually fits the workflow.
Use your current agents and harness as the baseline.
Include prompting, supervision, review, and corrections.
Use the value of an hour returned to useful engineering work.
Incremental consumed credits, postpaid charges or direct-provider costs, plus runners. Exclude unused prepaid balance.
Include integration and workflow configuration time.
Use 1 to charge all onboarding effort to the first month.
A strong starting point for a measured rollout.
Capacity value calculated from your inputs. Includes consumed usage and onboarding; unused prepaid credits are not an extra expense here. Applicable taxes are additional.
Net value = tasks × (current human minutes − new human minutes) ÷ 60 × hourly value − $600 − incremental usage − allocated onboarding cost.
Official list prices checked November 4, 2026. Different packages solve different parts of the problem.
| Option | Published monthly price | What the price buys | What to account for |
|---|---|---|---|
| OpenHands Agent Canvas | Free | Open-source, self-hosted agent operation and automations. | Your model usage, infrastructure, setup, and upkeep. |
| OpenHands Teams | $500 / team | Up to 10 users; self-hosted controls, integrations, and automations. | Own model keys. The published plan caps usage at 5,000 sessions/month. |
| Factory Max | $200 / individual | Coding agents with expanded usage allowances. | An agent subscription; compare your workload and current usage limits. |
| Devin Max | $200 / individual | Higher agent quotas and cloud-agent access. | Extra usage is separate. Team billing is a different package. |
| HarpessOps managed operations | $600 / team | Agreed workflows with bounded setup and recurring maintenance. | Prepaid model credits or postpaid at standard rates + 10%. Runner costs and custom work are separate. |
Agree on a starting baseline, put the workflows to work, and review the value with your team.
Record task volume, human minutes, accepted outcomes, and existing spend on representative tasks.
Track failures and review effort as carefully as successes. Include all incremental costs.
Compare net value with the alternatives. Use a 2× benefit-to-cost target to give the rollout room for operational variation.