Dock vs Multica: Action Permissions or Issue Review?

6 min read Comparisons

Dock sets permissions by agent action. Multica routes issues through project workflows. Compare control, memory and execution choices.

Dock and Multica put human control at different points: Dock lets you approve an agent's action type, while Multica hands an issue through project workflow steps. For a small team delegating recurring work, decide whether the person must authorize an action during execution or review the deliverable at a defined handoff.

Both offer cloud and local agent execution, people working alongside agents and recurring work from schedules or events. Those similarities will not settle the choice. Trace a real job from assignment to rejection or approval and identify which control must stop it from proceeding.

Dock coordinates specialists through a Chief of Staff

Dock's handoff guide describes a Chief of Staff briefing specialists, setting routines and keeping shared knowledge current. Handoffs transfer responsibility for finishing a task to an agent or person. Agents can escalate decisions up their reporting line.

People and agents work through shared docs, tables and channels. Dock's product page describes an append-only log recording who did what and when. That suits an operations team delegating research or sales work across specialists while retaining a record of the handoffs.

The Chief of Staff plans and coordinates through conversation, and shared documents can contain plans. Its public docs do not establish a conversation becoming a plan artifact reviewed before execution. Make that distinction when a stakeholder expects to approve the exact plan for a particular job.

Multica keeps the issue as the common record

Multica's issue documentation describes work assigned to a person, agent or squad. Discussions, deliverables, status and multiple agent runs stay on the issue. A project can have a person or agent lead, and squad leaders route issues to members.

Its project workflows give each step a next step and a send-back step for rejected work. A person can receive a review step, or a review agent can return work for revision. That fits teams accustomed to a shared issue board and explicit responsibility for each transition.

Multica also documents non-engineering work, including content, community work and recurring reports. The distinguishing model is the issue workflow, not a claim that only developers can use it. Its docs do not describe general conditional branches or per-requirement verifier checks in those project workflows.

Place permission before the risky action

Dock's permission guide gives each agent Allow, Ask or Deny settings by action type. Ask requires approval every time. Its Inbox collects work needing a decision and lets you answer without opening the workspace.

Multica's Inbox gathers notifications, and its workflows can assign a review step to a person. Its documentation does not describe configuring which agent actions need approval or collecting approval requests in that Inbox. A workflow handoff is published, while the same action-level permission model is not.

For a customer reply, these controls apply at different moments. Dock's permission settings can require a decision about an action type. Multica's project workflow can put a person at a review step and send rejected work back. Specify which moment must block your job before deciding that an Inbox or review stage satisfies the requirement.

The result checks also deserve a narrow reading. Dock says agents check their own results and the Chief of Staff performs a nightly self-review. Multica supports issue acceptance criteria and rejected-work handoffs. Neither set of docs establishes evaluation of delivered work against defined criteria or tests in the same way as an explicit verifier gate.

Compare recurring runs and persistent context

Dock's routine documentation supports hourly, daily, weekly and one-off routines in your timezone. Webhook rules can wake particular agents, with optional batching between check-ins. A local agent needs the Mac app open, and a Mac waking more than 15 minutes after a routine was due causes that occurrence to be skipped.

Multica's Autopilot guide describes schedules and webhooks that can create an issue each time or run without one. Issue wakeups can return agents to existing issues, although regular checks require an end date. Run only Autopilots skip runs when the runtime is unavailable and do not automatically retry failed runs in that mode.

Memory presents another concrete difference. Dock's Memory guide lets you read, edit or delete entries that persist into later conversations. Multica retains Hermes memory per agent and can optionally enable Codex native memory, but its documentation does not describe user inspection or correction. If a recurring report uses an outdated instruction, establish how you will repair its saved context before depending on the next run.

Separate agent location from service ownership

Dock's cloud-agent guide offers cloud agents on Dock's machines and local agents through its Mac app. Cloud agents continue with the laptop shut. A connected Claude subscription can fund local agents, while cloud agents always use Dock usage. Not every OpenAI model supports subscription funding.

Multica's deployment documentation permits self-hosting the service. It also supports a connected laptop or server running installed coding tools, alongside managed cloud sandboxes. Those tools can use your local sign-in or model-provider setup, with model availability determined by the runtime and account entitlements.

Dock's pricing page lists Starter at $40, Standard at $100 and Scale at $200 per person per month. Plans include 5, 10 or 20 cloud agents and 10,000, 20,000 or 50,000 business credits respectively. Dock is in beta, and its terms do not offer a formal uptime service-level agreement.

Multica's homepage says internal organizational use, including self-hosting, is free. External hosting or embedding requires a commercial license, and the page does not publish a cloud price. Treat the license and the execution arrangement as separate questions when planning a client deployment.

Dock and Multica, point by point

The following table compares both products against the same questions. Each answer includes its source and check date so the narrower details remain visible alongside the label.

Both columns answer the same questions from each company's public pages. Yes means the company publishes the capability the question asks about, Partly means something narrower, No means it says it does not offer it, and Not published means its public pages do not describe it. Every answer links to its source and shows when we last checked it. The oldest check on this page was November 13, 2026.

Dock and Multica compared, with the source and check date of each answer
What you compare Dock Multica
Goals Dock Partly

Agents accept goals and delegate parts to specialists, and a channel example shows progress toward a lead target. Its docs do not describe tracked goal metrics beyond conversational progress reports.

trydock.ai/product
Multica Partly

Projects have a shared goal and a member or agent lead. The documentation does not describe linked outcome metrics monitored by that lead.

multica.ai/docs/projects
Leads Dock Yes

A Chief of Staff runs work involving several agents, briefs each specialist, sets routines and keeps shared knowledge current. Agents escalate decisions up their reporting line.

trydock.ai/help/hand-work-off
Multica Yes

Project leads can be people or agents, and squad leaders route issues to members. The documentation does not describe goal leads that monitor outcomes or project leads that independently review goal conflicts.

Tasks Dock Yes

Handoffs transfer responsibility for finishing a task to an agent or a person. An append-only log records who did what and when.

Multica Yes

Assign an issue to a person, agent or squad. Its discussion, deliverables, status and multiple agent runs remain on the issue.

multica.ai/docs/issues
Recurring work Dock Yes

Agents run hourly, daily, weekly or one-off routines in your timezone. Webhook rules wake specific agents from outside events, with optional batching between check-ins.

Multica Yes

Autopilots start from schedules or webhooks, creating an issue each time or running without one. Recurring wakeups also return an agent to an existing issue, but regular checks require an end date.

Workflows Dock Partly

You can split recurring work among specialists and specify handoffs and human decisions. Dock describes agents adapting their steps, rather than a defined multi-step process that runs the same way each time.

trydock.ai/blog/agentic-workflows
Multica Yes

Project workflows hand steps to agents or people, with a next step and a send-back step for rejected work. They do not document general conditional branches or per-requirement verifier checks.

multica.ai/docs/workflows
Autonomy Dock Yes

Each agent has Allow, Ask or Deny permissions by action type, with Ask requiring approval every time. The Inbox collects work needing a decision and lets you reply without opening the workspace.

Multica Partly

An Inbox collects notifications and workflows can hand a review step to a person. The documentation does not describe configuring which agent actions require approval or collecting approval requests in the Inbox.

Learned autonomy Dock Not published

Its docs describe manually moving permissions from Ask to Allow and learning your working preferences. They do not describe proposals to change an agent's authority based on your approval history.

Multica Not published

Multica documents agent access and workflow handoffs. Its documentation does not describe proposing changes to an agent's autonomy based on approval history.

Memory Dock Yes

Agent memory updates during work and carries standing knowledge into later conversations. You can read, edit or delete entries in the Memory tab.

trydock.ai/help/teammate-memory
Multica Partly

Multica retains Hermes memory per agent and can optionally enable Codex native memory. Its documentation does not describe whether users can inspect or correct that memory.

multica.ai/docs/daemon-runtimes
Planning Dock Partly

The Chief of Staff plans and coordinates work in conversation, and shared docs can hold plans. Its docs do not establish a conversation becoming a plan artifact reviewed before execution.

Multica Yes

Chat agents can create issues when asked, and issues can contain acceptance criteria. The documentation does not describe a required plan and acceptance-criteria review before an agent starts.

Intake Dock Yes

Each connected tool gets a webhook URL, and plain-language rules route events to agents. Channel events can also tag the right person and let an agent pick up work in the thread.

trydock.ai/help/webhooks
Multica Yes

A webhook can trigger an autopilot that creates an issue, and a squad leader can route an assigned issue to members. The documentation does not describe a project-token intake that places a proposal in the Inbox first.

Connectors Dock Yes

The official CLI documents MCP endpoints, API keys and webhook configuration. Incoming webhook rules wake agents from connected services.

Multica Yes

Agents can use individual or shared MCP servers, and autopilots accept webhook triggers. The documentation does not describe the same catalog of API, command-line and MCP connectors.

Models Dock Partly

You choose a model and reasoning effort for each agent and can switch models in the composer. Its docs do not describe automatic model selection for each task or step.

trydock.ai/help/choose-a-model
Multica Partly

You can select a model and thinking level for supported agent runtimes. The documentation does not describe automatic model selection for each work stage based on price and budget.

multica.ai/docs/agents-create
Cloud or your own computer Dock Yes

Choose cloud agents on Dock's machines or local agents through the Mac app. Cloud agents keep working with your laptop shut, while local agents can access your Mac's files and tools.

trydock.ai/help/cloud-agents
Multica Yes

Connect a laptop or server to run installed coding agents, or use Multica's managed cloud sandboxes. You can also self-host the Multica service.

Your own subscription Dock Partly

Connected Claude subscriptions and supported ChatGPT models can fund local agents instead of Dock usage. Cloud agents always use Dock usage, and not every OpenAI model supports subscription funding.

Multica Yes

Multica invokes installed AI coding tools such as Claude Code, Codex and GitHub Copilot CLI using your local sign-in or model-provider setup. Available models depend on the tool, local configuration, signed-in account and subscription entitlements.

Pricing Dock Yes

Starter is $40, Standard $100 and Scale $200 per person per month, including 5, 10 or 20 cloud agents and 10,000, 20,000 or 50,000 business credits. Additional usage or credits can be purchased and roll over.

Multica Partly

Multica says internal organizational use, including self-hosting, is free, and external hosting or embedding requires a commercial license. It does not publish a cloud price on its home page.

multica.ai
Correctness Dock Partly

Dock says agents check their own results, and the Chief of Staff runs a nightly self-review to retain lessons. Its docs do not describe evaluating delivered work against defined criteria or tests.

Multica Partly

Issues can contain acceptance criteria, and a workflow can send rejected work back from a review step. The documentation does not establish that results are checked against defined criteria.

A fact we cannot find or cannot check is left out, and prices can change after the date shown. Task Machine publishes this comparison and is an interested party. Dock and Multica are named only to identify them, and there is no relationship between the companies. Has something changed? Tell us.

Who each one suits

Pick Dock if you want a Chief of Staff coordinating business specialists with per-agent action permissions and editable memory. Account for its beta status and the local routine skip rule before giving it time-sensitive recurring work.

Pick Multica if people and agents should share issues with project review steps and you want the option to self-host. Its published control model does not establish per-action agent approvals, and self-hosting leaves backup and migration upkeep with your team.

Where Task Machine fits

Task Machine combines action settings with defined Workflows containing branches, human questions, approvals and verifier checks that return failed work with feedback. Tasks keep plans, decisions and runs together, and you review acceptance criteria before execution. Task Machine is the wrong pick if you require self-hosting the service, since its public setup docs do not describe self-hosting the workspace service. See Switch from Multica to Task Machine.

Compare Task Machine vs Dock and Task Machine vs Multica for that combination of controls. Next, map one recurring report's approval points and failure path, including what should happen when the local computer misses its scheduled run.