Paper Instruments

Grok Bot alternatives, from standing bots to professional agents

This guide weighs Ultramarine Work, Claude Cowork, ChatGPT Work, Perplexity Computer, Manus, Genspark, Lindy, Viktor and OpenClaw against Grok Bot, covering what each is built for, which models run it and what access costs. A worked Sales assignment then follows one renewal from source files to the signed letter.

Checked on

At a glance

ProductBest forAvailable onStarts atFree access
Grok BotNamed Bots for standing operations workDesktop, mobile, X, Slack, cloud computerIncluded with X Premium, $30 a month7-day trial usage credit
Claude CoworkDelegated file work and scheduled tasksDesktop, web and mobile betaIncluded from Claude Pro, $20 a monthNeeds a paid Claude plan
ChatGPT WorkLong multi-tool tasks that end in filesDesktop, web and mobile for paid plansPlus $20 a month, Pro from $100Limited, desktop app only
Perplexity ComputerResearch-heavy work with cited sourcesWeb, mobile, Comet, Mac appPro $20 a month, Max $200Limited free turns on some accounts
ManusAutonomous builds like sites and decksWeb, desktop, mobile, cloud computerPro from $20 a month300 daily credits, one task
Ultramarine WorkProfessional modes with curated skillsWeb and desktop for Mac and Windows$5 a month, after 7 days free7 days free on Starter and Plus
Genspark (including Claw)One prompt into decks, sheets and sitesWeb, mobile, Claw in messaging appsPlus $24.99 a month100 credits a day, lifetime cap
LindyInbox, meetings and CRM upkeepSlack, email, phone, computer useFrom $49.99 a user a month$50 in credits for 7 days
ViktorOne shared coworker in your chat toolsSlack, Teams, Discord, webTeam from $50 a monthUp to $100 in credits that never expire
OpenClawSelf-hosting with your own modelYour own hardware, via messagingFree software, you pay model costsFree, open-source software

What people use Grok Bot for

Teams give Grok Bot standing operations work: SpaceXSI’s sales team runs Account Intel Bot for pre-call research, Quote Bot for quotes and Deal Bot for pipeline hygiene. Each Bot carries its own role on a persistent cloud computer and saves a demonstrated workflow as a routine.

Grok Bot alternatives

  1. 1. Claude Cowork

    Anthropic

    Good for
    Delegating real work to one assistant: organizing and editing files in a folder, producing reports and decks, and running scheduled tasks while your laptop is closed. The marketplace lists 38 plugins, including Anthropic’s role plugins for sales, finance and legal, and plugins are plain files.
    How it works
    One Claude agent on the Claude Code harness, extended by opt-in plugins bundling skills and connectors.
    Runs on
    Anthropic’s own Claude models (Opus, Sonnet, Haiku, Fable). Model and agent come from the same company.
    Price
    Included from Pro, $20 a month or $17 billed annually. Max $100, Ultra $200, Team seats from $20. Ultra has no 5-hour limit for verified accounts.
    Choose it when
    Your team already drafts and analyzes in Claude and wants one assistant moving between writing, folder cleanup and scheduled jobs, with its process in plugin files it owns.

    Sources: Anthropic, Claude Cowork product page, Anthropic, Claude pricing, Anthropic support article on Cowork

  2. 2. ChatGPT Work

    OpenAI

    Good for
    Hours-long, multi-tool assignments that end in a file: pull CRM data, gather Slack context, then produce a spreadsheet, deck or small web app, pausing for approval on consequential actions. It has a wide plugin catalog and the largest consumer user base of any assistant.
    How it works
    An agent for long tasks inside ChatGPT, alongside Chat and Codex in the redesigned desktop app.
    Runs on
    OpenAI’s own models: GPT-5.6 (Sol, Terra, Luna) at launch, with GPT-6.1 Sol rolling out into Work from late September 2026.
    Price
    Free and Go ($8) include limited Work, desktop app only. Plus $20 adds web and mobile, Pro runs $100 to $500, Business $25 a user monthly.
    Choose it when
    The work already runs through ChatGPT: the desktop app is free to try, Pro runs several agents in parallel, and plugins cover Slack, Teams, Salesforce and Canva.

    Sources: OpenAI, ChatGPT Work and Codex help center, OpenAI, Learn pricing page, OpenAI, ChatGPT pricing page

  3. 3. Perplexity Computer

    Perplexity

    Good for
    Research-heavy work where cited sources matter: an outcome is split into tasks, run by sub-agents in cloud sandboxes, and returned as documents, spreadsheets, dashboards or apps. Reusable Skills keep recurring workflows consistent, and Personal Computer keeps the agent working on your own Mac.
    How it works
    A general-purpose digital worker that breaks a goal into sub-agents, plus an always-on option on your own Mac.
    Runs on
    Third-party models on Perplexity’s harness: Claude Opus 4.6 drove reasoning at launch (February 2026), with GPT-5.6 defaults since, per secondary changelogs.
    Price
    Pro, $20 a month, adds a weekly Computer allowance still rolling out and a one-time 4,000-credit bonus. Max, $200, brings 10,000 credits at 100 credits to $1.
    Choose it when
    The output should carry citations, you want published per-task credit examples before you commit, or an agent should keep working on a Mac that never sleeps.

    Sources: Perplexity, Computer launch post, Perplexity, how credits work

  4. 4. Manus

    Manus

    Good for
    Hands-off deliverables from a plain-English goal: websites, slide decks, wide research over many parallel tasks, video ads and simple games. Paid plans run 20 concurrent tasks, the Manus Studio desktop covers docs, sheets, slides, sites and code, and automations plus a purchasable cloud computer keep projects moving.
    How it works
    An autonomous agent that plans and executes in a cloud virtual machine. Its Cue app gives personal agents their own phone and wallet, but it is invite-only.
    Runs on
    Core model not disclosed. Its own Cascade harness shipped with Manus 2.0 (September 2026). In March 2025 it said it used Claude with fine-tuned Qwen models.
    Price
    Free plan: 300 credits a day and one concurrent task. Pro starts at $20 a month for 4,000 credits, or $40 for 8,000. Team seats start at $20 a month.
    Choose it when
    You want one-shot builds, such as a finished site, deck or video, and the 300 free daily credits and 20 parallel tasks give room to try them.

    Sources: Manus, Manus 2.0 announcement, Manus, membership pricing help article

  5. 5. Ultramarine Work

    Paper Instruments

    Ultramarine Work is a tool for a human, not a human-shaped tool: you stay the professional, and the mode changes the agent for your work. It fits account executives who hand over real assignments and expect reviewable files back: renewals, quotes and deal desk approvals. A mode is set once for a workstream and stays selected, and the Sales mode carries deal desk review and price increase rollout skills. For standing bots on routine operations, a personified agent fits better.

    Every mode pairs a curated skill library with Ultramarine 1 Flash, our model, post-trained on GLM-5.3-Flash for knowledge work. Because Paper Instruments writes the skills and trains the model, improvements to methods and model land together. You choose the way of working per task: Standard runs Ultramarine 1 Flash, Assisted adds a frontier model as an advisor, and Advanced hands the whole task to a frontier model. The skills carry the profession’s methods, review standards and common traps, and the person reviews every file before it goes out.

  6. 6. Genspark (including Claw)

    MainFunc Inc.

    Good for
    Turning a single prompt into a polished deliverable: slides, sheets, docs, websites, dashboards, images and video. Claw, the closest analog to Grok Bot here, is an agent on a dedicated always-on cloud computer you message through WhatsApp, Telegram, Teams or Slack.
    How it works
    A tool built around a Super Agent. Claw is framed as one AI employee per user but carries no role personas, unlike Grok Bot’s named Bots.
    Runs on
    Third-party models: Genspark says it orchestrates more than 70 models in a mixture-of-agents design, naming partnerships with OpenAI and Anthropic.
    Price
    Free: 100 credits a day, stopping once a lifetime allowance is used. Plus $24.99 a month ($19.99 annual) for 10,000 credits, Pro $249.99.
    Choose it when
    The finished deck or visual in hand matters most, or a Claw agent reachable from your messaging apps matches how you already delegate.

    Sources: Genspark, membership plans help center, Genspark, credits guide

  7. 7. Lindy

    Lindy

    Good for
    Admin on repeat: inbox triage and drafted replies in your voice, meeting recording and notes, scheduling and prep, CRM updates, support queues and recruiting outreach. It brings more than 40 built-in skills, 3,000-plus integrations, computer use and scheduled routines, and lives in Slack and Gmail.
    How it works
    A named teammate you hire for a role, such as account management, that works in your email, calendar and Slack. Actions with outside impact wait for approval.
    Runs on
    Third-party and model-agnostic: you pick the model for each task, and most major models are available.
    Price
    Free: $50 in credits for 7 days, no card required. Team from $49.99 a user a month for 5,000 credits, topped up at $10 per 1,000 credits.
    Choose it when
    Your delegation already lives in email, Slack and calendars and repeats on schedule, with role-based teammates, human approval gates and pooled team credits.

    Sources: Lindy pricing page

  8. 8. Viktor

    Viktor

    Good for
    One @-mentionable coworker for the whole team: it connects to more than 9,800 tools, runs recurring automations, builds internal tools, dashboards and apps, and hands results back in Slack, Teams or email. Credits are shared workspace-wide with no per-seat license, and unused plan credits roll over one month.
    How it works
    A single named agent your team @-mentions in Slack or Teams.
    Runs on
    Third-party models passed through at no markup: credits track what Anthropic, OpenAI and others charge.
    Price
    Free: up to $100 in credits that never expire. Team from $50 a month for 20,000 credits shared across the workspace, with larger tiers above.
    Choose it when
    Your team wants to share one agent from Slack without seat licenses, turn requests into small internal tools, and pay credits that track model prices.

    Sources: Viktor pricing page

  9. 9. OpenClaw

    OpenClaw Foundation

    Good for
    Teams with technical hands that want a personal agent on their own hardware, instructed through the messaging apps they already use. You can read and modify the source, audit what it does, choose any model backend, and keep logs inside your own infrastructure, under the independent OpenClaw Foundation.
    How it works
    A self-hosted, open-source agent that runs where you install it and talks through your existing messaging apps.
    Runs on
    Bring your own model: it supports multiple model backends, including open-weight models.
    Price
    The software is free under an open-source license. You pay for model access and hosting separately, so the true cost depends on your usage and infrastructure.
    Choose it when
    Data has to stay on infrastructure you control, you want to choose the model yourself, or you want to change the agent itself. It asks for setup time no hosted product demands.

    Sources: OpenClaw Foundation, Computerworld, OpenClaw coverage

Pricing and how to try it

Plans start at $5 a month, and Starter and Plus begin with 7 days free. On our internal PI Knowledge Work benchmark, Ultramarine 1 Flash costs about $0.13 a task, about a tenth of Claude Opus 5.5 or GPT-6 Sol at comparable scores.

Start free trial See every plan

Questions

Is Grok Bot free to use?

No, it is never sold on its own. Grok Bot comes with The X Subscription Premium at $30 a month and above, or with a legacy paid Cursor plan, and it is excluded from Free and Plus subscriptions. Its usage draws from the same weekly pool as Cursor and the Grok apps, and new accounts can get a 7-day trial usage credit.

Which alternatives can I try without paying?

Manus gives 300 daily credits on a free plan, Genspark gives 100 a day up to a lifetime allowance, Viktor fronts $100 of credits that never expire, and Lindy gives $50 of credits for a week, no card required. ChatGPT Work is limited but usable on desktop for free. Ultramarine Work has no free tier, and its Starter and Plus plans begin with 7 days free.

Can any of these act without approval?

Every product here keeps the person in charge, but the gates differ. Grok Bot waits for approval before consequential actions. Lindy holds anything with outside impact for approval. Ultramarine Work never signs, files, sends or posts to a system of record, and the person reviews every file.

Will I know which model answered in Grok Bot?

There is no way to see it directly. Grok Bot picks the model for each task and has no model picker: it runs SpaceXSI’s own models or third-party models from providers on its sub-processor list, you cannot see which one ran, and team model allowlists do not apply to it.

Do I have to choose one product for everything?

No, the division of labor keeps them apart. Standing bots from Grok Bot, Lindy or Viktor keep routine operations moving in your tools, while a coworking tool such as Claude Cowork or Ultramarine Work carries the assignments that end in reviewable files. The renewal above is the second kind of work.

Which options work for self-hosting and full data control?

OpenClaw is the open-source route: it runs on your own hardware, talks through your messaging apps, and leaves the model and hosting to you, which means setup and maintenance are yours. Everything else on this page is hosted.

Sources

Checked on