Roma vs Tasklet: Shared Lists or Shared Agents?
Roma and Tasklet both connect people with agents. Compare task records, automation planning, memory and execution before choosing.
Founder, Task Machine
Roma and Tasklet put collaboration in different places: Roma keeps people and agents on shared task records, while Tasklet lets teammates work with shared agents in threads. When a colleague takes over an unfinished job, the deciding question is where they will find the assignment, discussion and result.
Both describe recurring automations and tool permissions. Roma also publishes a process for approving an automation's schedule and steps, while Tasklet describes shared skills and review of those reusable instructions. That difference matters when the team needs to agree on how a recurring brief will run before enabling it.
Roma makes the task the meeting place
Roma's shared-list guide separates Assigned to, which names a person, from Delegate, which hands work to the Operator or a connected agent. The task body holds work, the conversation holds discussion, and the Log records edits. Teammates can read a shared task's conversation and reply to its connected agent.
That arrangement fits a team that already thinks in lists of work. A person can remain responsible for an item while asking an agent to contribute, and another person can find the work on that same record. The task is the place to inspect a handoff.
Roma also supports Mac and iPhone capture and offline list edits. The Operator requires the network, so offline editing should not be confused with offline agent execution. Roma's public docs do not describe an owned measurable goal that an agent monitors, making a task list and an outcome-monitoring system separate buying questions.
Tasklet makes the agent reusable across threads
Tasklet's agent guide describes agents with instructions, connections, knowledge and automations. Teammates can share an agent and open separate threads for work. An agent-wide memory carries context between threads, which fits a shared assistant that repeatedly handles related requests.
The published scope differs from Roma's assigned task record. Tasklet agents can use a task list during multi-step work, but its docs do not describe a shared task record owned by either a person or an agent. Inspect how your team would distinguish a colleague's responsibility from the agent's current thread.
Sharing also sets an access boundary. Tasklet's guide says members of a shared agent can see and use every thread in that agent, and recommends separate agents when access should differ. A thread split alone should not be your plan for separating client work.
Agreeing on the recurring job
Roma's automation creation guide describes a conversation with the Operator, tool rehearsal and a proposed card containing the schedule and steps. You can request changes before approving the card and enabling the automation. Its automations can follow a numbered playbook and retain a result for each run.
Tasklet publishes scheduled and event-driven work, including Gmail, Slack and webhook triggers. Its shared-drives article describes asking an agent to identify gaps and draft a shared skill for review. Its docs do not establish a repeatable multi-step process that runs the same way each time, or a reviewed execution plan for a particular job.
For a weekly support brief, review a concrete artifact during evaluation. In Roma, inspect the proposed schedule and playbook before enabling it. In Tasklet, inspect the skill and trigger configuration, then decide whether those instructions provide the repeatability you require.
Roma's custom MCP tools have a specific automation boundary. Scheduled automations can use a custom server's reading tools only, while tools that change or remove something are available in chat. Verify that boundary if the brief must also update another system.
Correcting context and controlling tools
Roma's memory guide describes an account-level journal and profile built from tasks, conversations and connected apps. You can read the profile and correct it on the Memory page, and later replies use that correction. This gives you a documented place to repair an incorrect preference.
Tasklet's memory announcement describes context that follows an agent across threads, including conversations and files. Its docs do not describe a way to inspect or correct that memory. The guide also says memory stays enabled once switched on for an existing agent.
Both products publish tool controls. Roma offers Always allow, Ask first and Never for Operator actions, with approval cards in chat. Tasklet sets tool permissions per agent and asks when a tool is not enabled. Neither product's docs describe proposing permission changes from your approval history, so plan for a person to revisit those settings.
Choosing the execution and billing boundary
Roma runs Operator automations on its servers, while external agents can listen for task events on a laptop or server. Its documented Claude delegation uses your paid Claude plan for the connected agent's work, not the built-in Operator. Its pricing address does not publish service plan prices or included model usage, so the monthly service cost remains unverified.
Tasklet's pricing page lists limited free trial credits and a $25 monthly plan with 10,000 credits shared across the organization. It also offers one-time credits. Agents run in managed cloud sandboxes, and its docs do not describe connecting your own AI subscription or model-provider API key.
Tasklet's Auto can select a model per message with Economy, Balanced or Premium settings. Roma's developer guide says it cannot set a connected cloud agent's model or reasoning. That favors different decisions: delegating to an agent you already chose, or asking the hosted agent service to select a model.
Roma and Tasklet, point by point
The table compares their answers to the same questions. Each explanation includes its source and check date, so similar labels can be read in context.
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.
| What you compare | Roma | Tasklet |
|---|---|---|
| Goals |
Roma
Not published
Roma documents tasks and lists with custom properties. Its public docs do not describe an owned goal with a measurable target that an agent monitors and acts on. |
Tasklet
Partly
An agent can report on revenue, support tickets and project milestones each morning. Its docs do not describe an agent owning a goal with a measurable target. |
| Leads |
Roma
Not published
Shared lists have owners, and tasks can name a person or delegate. Its docs do not describe lead roles that review and coordinate a project, goal or team. |
Tasklet
Partly
An agent can assign sales accounts using your rules and delegate focused work to subagents. Its docs do not describe a named lead coordinating a project, goal or team. |
| Tasks |
Roma
Yes
Assigned to names a person, while Delegate hands a task to the Operator or a connected agent. Each task keeps its own conversation, work in its body and a Log of edits. |
Tasklet
Partly
Agents use a task list for multi-step work, and teammates can work with shared agents in threads. Its docs describe no shared task record owned by either a person or an agent. |
| Recurring work |
Roma
Yes
Automations run on schedules or after completion. List changes and delegation events wake connected agents through polling or webhooks, including changes made through the API. |
Tasklet
Yes
Agents run on schedules and events, including Gmail, Slack and webhooks. Its docs describe no stable root task with separate occurrence histories. |
| Workflows |
Roma
Yes
An automation can follow a numbered playbook on every run, such as reading issues, checking mail and writing a brief. Each run keeps its result, and you approve the steps before switching it on. |
Tasklet
Partly
Shared skills describe repeatable work, and agents can request approval by email. Its docs do not establish a repeatable multi-step process that runs the same way each time. |
| Autonomy |
Roma
Yes
Operator actions can be Always allow, Ask first or Never. Approval cards appear in chat, and automations have their own per-app and per-action permissions within your connection settings. |
Tasklet
Yes
Tool permissions are set separately for each agent, and an agent asks before using a tool that is not enabled. You can also decide which drafts require approval. |
| Learned autonomy |
Roma
Not published
You can change permissions or choose Always allow on an approval card. Its docs do not describe Roma proposing permission changes from your approval history. |
Tasklet
Not published
Tasklet lets you configure tool permissions for each agent. Its docs describe no proposals to change permissions based on approval history. |
| Memory |
Roma
Yes
Roma builds an account-level journal and profile from tasks, conversations and connected apps. You can read the profile and correct it on the Memory page, and later replies use the correction. |
Tasklet
Partly
Agent-wide memory carries context across threads, including conversations and files. Its docs do not describe a way to inspect or correct the agent's memory. |
| Planning |
Roma
Yes
You can discuss an automation with the Operator, which rehearses its tools and proposes a card with the schedule and steps. You can request changes before approving the card and enabling the automation. |
Tasklet
Partly
You can ask an agent to identify gaps and draft a shared skill for your review. Its docs do not establish a reviewed plan for a particular job before execution. |
| Intake |
Roma
Yes
Forwarded email from trusted senders becomes a task on your account's Tasks list. The REST API also creates items on a specified list, and tasks can then be delegated to an agent. |
Tasklet
Yes
An agent can receive leads through a webhook and distribute them round-robin to sales representatives. |
| Connectors |
Roma
Yes
Roma accepts custom remote Model Context Protocol servers, publishes a REST API and sends filtered list-change events to agents by webhook or polling. Automation runs can use custom servers' reading tools only. |
Tasklet
Yes
Tasklet connects agents to over 3,000 built-in integrations, HTTP APIs and MCP servers, and provides a sandboxed command line. Its docs describe no custom command-line connector catalog. |
| Models |
Roma
Not published
Roma connects to external agents, and its developer guide says it cannot set a cloud agent's model or reasoning. Its docs do not describe automatic model selection per task or step by Roma. |
Tasklet
Yes
Auto selects a model for each message with Economy, Balanced or Premium settings, and you can choose a specific model or an agent default. Agents can delegate parts of work to other models. |
| Cloud or your own computer |
Roma
Yes
Operator automations run on Roma's servers. Connected agents can instead listen for work on your laptop or server through the documented polling loop. The own-computer option uses your external agent, not a local Operator. |
Tasklet
Partly
Tasklet agents run in managed cloud sandboxes, with a Mac app for access. Its docs do not describe running agents on your own computer or infrastructure. |
| Your own subscription |
Roma
Partly
Roma can delegate tasks to Claude using your paid Claude plan. This covers the connected agent's work, not Roma's Operator, for which its docs describe no subscription connection or provider-key option. |
Tasklet
Not published
Tasklet provides models in one plan and charges Tasklet credits for agent work, including work delegated from other AI apps. Its docs do not describe using your own AI subscription or model-provider API key. |
| Pricing |
Roma
Not published
The US App Store lists the iPhone download as Free. Roma's reviewed public pages do not publish service plan prices, included model usage or usage rates, so a monthly service cost is unverified. |
Tasklet
Yes
Tasklet offers limited free trial credits and a $25 monthly plan with 10,000 credits shared across the organization. One-time credits are also available. |
| Correctness |
Roma
Partly
Connected agents leave work and updates on the task, and their instructions leave completion to you. Its docs describe no result checks against defined acceptance criteria, rubrics or tests. |
Tasklet
Partly
A customer describes an agent checking live ad settings after applying approved changes, and another agent checking ads against placement and audience rules. |
For software work
| What you compare | Roma | Tasklet |
|---|---|---|
| Code review rules |
Roma
Not published
The GitHub Operator reads a repository's root AGENTS.md and follows its conventions when writing files. Its docs do not describe reviewing code changes against repository rules. |
Tasklet
Not published
Tasklet agents can inspect repository history and work with pull-request reviews. Its docs describe no review against repository rules. |
| Adversarial review |
Roma
Not published
Roma offers a Devil's Advocate expert for discussing ideas. Its docs do not describe an independent second review of risky code changes. |
Tasklet
Not published
Tasklet agents can work with pull-request reviews and discussions. Its docs describe no independent second review for risky changes. |
| Risk rating |
Roma
Not published
The GitHub Operator can open and merge pull requests when asked, with approval cards for those actions. Its docs do not describe a risk rating before merge. |
Tasklet
Not published
Tasklet agents can inspect changed pull-request files and, with permission, merge pull requests. Its docs describe no merge risk rating. |
| Pull requests |
Roma
Partly
The GitHub Operator opens file writes as pull requests and can read reviews and update pull requests. Its docs do not establish an agent loop addressing review feedback, and the Operator runs no builds or tests. |
Tasklet
Yes
Tasklet can create pull requests, work with reviews and discussions, and merge with permission. Its docs describe no recorded approval of the exact commit before merging. |
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. Roma and Tasklet are named only to identify them, and there is no relationship between the companies. Has something changed? Tell us.
Who each one suits
Pick Roma if shared lists, explicit human assignment and agent delegation are the foundation of your team's work. Stay aware that its service prices are unverified and scheduled custom MCP tools are limited to reading.
Pick Tasklet if shared agents, connected tools and per-message model selection fit your team's requests. Its public docs do not describe an inspectable memory editor or a human-or-agent-owned shared task record.
Where Task Machine fits
Task Machine brings plan review, ownership and acceptance checks into each Task. Its task-review documentation describes checking delivered work against the requirements before a person reviews it, and Workflows can return failed steps with feedback. That addresses a different requirement from sharing an assistant or approving a scheduled playbook.
Task Machine is not the right fit when your main requirement is Roma's offline list editing or Tasklet's shared-agent thread model. Compare Task Machine vs Roma and Task Machine vs Tasklet, then choose one unfinished team job and verify where its owner, approved plan and final review would live.