Remotasks sits in an interesting position in the AI training data market. It is backed by Scale AI — one of the most well-funded and influential AI data companies in the world, with contracts across major technology companies and defense organizations — which immediately puts it on a different footing from the scrappier crowdsourcing platforms that appear and disappear every few years. But backing from a credible company does not automatically translate into a good experience for contributors. This review covers what Remotasks actually is, what the work looks like in practice, what you can realistically earn, and who the platform genuinely suits in 2026.
What Is Remotasks and Who Runs It
Remotasks is the crowdsourcing division of Scale AI, a San Francisco-based data company founded in 2016 that has become one of the most important players in the AI training data industry. Scale AI provides the labeled datasets that major technology companies, autonomous vehicle developers, and AI research labs use to train their models. To produce this data at scale, the company operates Remotasks as the platform through which independent contributors around the world complete annotation and labeling tasks remotely.
This corporate structure matters for understanding what Remotasks is and is not. It is not an independent crowdsourcing platform — it is a direct pipeline into Scale AI's client work. The tasks you complete on Remotasks feed directly into the training datasets Scale AI delivers to its clients. That gives the platform a level of continuity and funding that purely independent crowdsourcing platforms lack, but it also means that task availability on Remotasks tracks directly with Scale AI's client contracts, which can shift quickly.
Getting started on Remotasks is free and requires only a basic account registration. Unlike Appen or Telus International AI, there is no invitation-only waitlist — you can sign up and begin working through the onboarding process immediately. The platform is available globally, though task availability and pay rates vary significantly by region and project.
What Tasks Are Available on Remotasks
Remotasks offers a wider variety of task types than most AI training platforms, and the available work has expanded considerably as Scale AI's client portfolio has grown. The main task categories include:
Image annotation — the most widely available task type. This includes drawing bounding boxes around objects in photos, marking specific attributes of those objects, and classifying images according to defined categories. Image annotation on Remotasks ranges from basic labeling (accessible to new contributors) to highly precise semantic segmentation (requiring specific training and accuracy standards).
3D point cloud annotation — labeling three-dimensional lidar scan data, primarily for autonomous vehicle development. This is the task category that most distinguishes Remotasks from general crowdsourcing platforms. Lidar data looks like a three-dimensional cloud of points representing the physical environment scanned by a sensor mounted on a vehicle. Annotators identify and label objects — vehicles, pedestrians, cyclists, road infrastructure — within this 3D space. The work requires significant training to do well, but pays substantially more than basic 2D annotation tasks.
Video annotation — tracking objects across video frames, ensuring that labels applied to moving objects remain accurate throughout a sequence. More demanding than single-image annotation but generally more engaging and better paid.
Natural language processing tasks — reviewing, correcting, or rating text and AI-generated responses for quality, accuracy, and safety. This category has grown significantly in recent years as demand for RLHF (Reinforcement Learning from Human Feedback) training data has expanded.
Audio and speech tasks — transcribing audio, identifying speakers, classifying sounds, and rating the quality of speech data. These tend to be intermittent projects rather than consistently available work.
The Training Process: How to Access Better-Paying Work
One of the most important things to understand about Remotasks before you start is that the platform is structured around a training-first model. Most of the higher-paying tasks are not accessible immediately — you need to complete free training courses and pass accuracy assessments before being assigned paid work in those categories.
Remotasks offers its own training material through what it calls "Tasking Courses." These are structured lessons that walk you through the guidelines, tools, and annotation standards for a specific task type. Courses range from short introductions (completing in an hour or two) to longer technical programs covering complex annotation work like 3D lidar.
After completing a course, you take a qualification assessment — a set of practice tasks graded against the platform's accuracy standards. Passing the assessment unlocks access to paid tasks in that category. Failing means waiting and retrying.
The investment required is real. New contributors who jump in hoping to earn immediately often find that the first few hours are spent on unpaid training rather than paid tasks. This is frustrating if you came expecting immediate income, but the platform makes no secret of it — the training materials are free and comprehensive, and the pay rates for qualified work justify the upfront investment for contributors who complete the process.
Starting with simpler image annotation tasks first, building accuracy, and then progressing to more complex and better-paid categories is the most effective approach.
Pay Rates: What Remotasks Actually Pays
Pay on Remotasks varies more than on most platforms because the range between basic and advanced task categories is significant. Understanding where different task types fall on the pay scale helps set accurate expectations.
Basic image annotation tasks — these are the most accessible tasks and the entry point for most new contributors. Effective hourly rates for straightforward bounding box annotation typically fall between $3 and $8 per hour depending on task complexity and how quickly you develop speed. These rates are modest compared to other AI training platforms and are best treated as a starting point while you qualify for higher-paying categories.
Intermediate annotation tasks — tasks requiring attribute classification, more precise segmentation, or additional judgment calls pay better. Effective hourly rates in this range typically fall between $8 and $15 for contributors who have built accuracy and speed.
3D lidar annotation — this is where Remotasks distinguishes itself. Experienced contributors who have completed the lidar training program and developed proficiency in the platform's 3D annotation tools report effective hourly rates of $15 to $30 or more. Some specialized lidar projects pay at the high end of this range consistently for qualified annotators.
NLP and RLHF tasks — rates for text and AI output evaluation vary by project and are generally in the $10 to $20 per hour equivalent range for contributors who qualify.
Payments are processed weekly via PayPal or Wise, with no minimum payout threshold for most account types. The weekly payment schedule is a practical advantage over monthly-pay platforms, and Remotasks' payment history is generally reliable given Scale AI's financial backing.
The Remotasks Community and Discord
One aspect of Remotasks that sets it apart from most crowdsourcing platforms is the active contributor community, primarily organized through a Discord server. The Discord functions as a support channel, resource hub, and social space for contributors — and it is genuinely useful.
New contributors can find answers to training questions, get help understanding annotation guidelines, and learn which task types are currently active and worth pursuing. Experienced contributors share tips about accuracy improvement, tool shortcuts, and which training courses are worth prioritizing. The community is large enough to have real activity across multiple time zones.
Access to the Discord is typically available after account creation. For new contributors especially, spending time in the server early on — asking questions, reading pinned resources — accelerates the learning curve significantly compared to working through the training materials in isolation.
Who Remotasks Genuinely Suits
Remotasks is not the right starting point for everyone, and being clear about this upfront saves frustration.
It suits contributors who:
- Are willing to invest time in training before earning significant income
- Are interested in building a specific, higher-value skill (3D annotation in particular) rather than completing interchangeable micro-tasks
- Want their work connected to real AI development at a credible technology company
- Are comfortable with variable task availability and are using the platform as one of several income streams
It is less suited to contributors who:
- Need to start earning immediately without an upfront training investment
- Want consistently predictable daily task volume
- Are looking for the simplest possible entry into online earning
The platform's strongest differentiator is the ceiling — contributors who develop proficiency in 3D lidar annotation can earn at rates that most general crowdsourcing platforms cannot match. Getting there requires patience and deliberate skill development.
Honest Downsides Worth Knowing
Task availability is uneven. Like all crowdsourcing platforms tied to client contracts, Remotasks goes through busy periods and quiet periods. When Scale AI lands a major autonomous vehicle contract, lidar tasks can be abundant; when that project cycle ends, the queue can thin out significantly. This unpredictability is the most common complaint from established contributors.
Basic task pay is low. Entry-level annotation rates are below what Appen and Telus International AI pay for comparable work. The platform's value proposition is in the higher-paying specialized categories — if you only ever complete basic image tasks, the platform is not the best use of your time.
Account issues have been reported. A subset of contributors report sudden task access restrictions or account suspensions without clear explanation. These incidents appear to be connected to accuracy drops or guideline violations, but the communication around them is often poor. This is not unique to Remotasks but is worth noting as a risk.
Regional task variation. Contributors in certain regions consistently report better task availability than others. This reflects which client projects are active and which markets those projects focus on — something outside any contributor's control.
Is Remotasks Worth It in 2026
For contributors willing to treat it as a skill-building platform rather than a quick-start earning option, yes. The path from zero to meaningful income on Remotasks requires real investment — working through training, developing accuracy, qualifying for better task categories — but the pay ceiling for specialized annotation work is among the highest available in the crowdsourcing category.
The practical approach is to run Remotasks alongside other platforms rather than relying on it alone. During slow periods or while training is in progress, consistent earning from complementary sources fills the gap. NexGuild provides daily earning through tasks and surveys from CPX Research and TheoremReach at nexguild.in — no training required, and NexCoins redeem for gift vouchers immediately. The NexGuild community is also worth joining for platform tips alongside the Remotasks Discord.
For a broader comparison of where Remotasks fits against Appen, Telus International AI, and Clickworker, our best crowdsourcing platforms guide 2026 puts them side by side — and if you are building toward the search quality rating work that Appen and Telus specialize in, our Appen review and Telus International AI review cover both in detail.
Key Takeaways
- Remotasks is the crowdsourcing platform of Scale AI, one of the most significant AI training data companies — giving it funding stability most competitors lack
- Task types range from basic image annotation ($3–$8/hr) to 3D lidar annotation for autonomous vehicle development ($15–$30+/hr)
- Higher-paying work requires completing free training courses and passing accuracy assessments before accessing paid tasks
- Payments are processed weekly via PayPal or Wise with no minimum payout threshold
- Task availability fluctuates with Scale AI's client contracts — the platform works best as one of several income streams
- The active contributor Discord community is a genuine resource for new contributors working through training
- The platform rewards contributors who invest in developing specialized annotation skills rather than those looking for immediate, low-barrier earning
