Pricing

Our goal is to provide a simple, fair and cost-efficient pricing that mixes the ease & stability of a subscription with control & efficiency of usage-based billing - simple enough to start in seconds, flexible enough to stretch the spending in any needed direction.

This is why we offer the following subscription plans, each with specific monthly quota of TaskLift credits based on USD to cover typical usage depending on your involvement. You decide how to spend them with option to find out the current usage in the Usage section of your profile. Unused monthly credits are not transferred to the next billing period.

Model Pricing

This section describes how credits are used when using AI models depending on the model type. Learn more about model types on the  Working With Models page.

Built-In Models

We provide competitive prices for unified, zero-effort access to all major cloud model providers including OpenAI, Google, xAI, Anthropic and DeepSeek. Our built-in models often provide cheaper access, more models to choose from and faster response times than provider-side subscriptions.

Credits are spent based on the number of tokens used during chat with the model. Rates per 1M input & output tokens are model-specific and are specified on each built-in provider's page.

In practice, the Starter subscription is enough to cover millions of leading model tokens.

Private Models

We can host your own private models with a wide choice of models supported by Ollama that fit the NVIDIA L40S GPU (models up to 40B parameters are usually fine). We offer very competitive prices while also covering all the hosting-related heavylifting.

Private model usage is charged by hosting time of the server that we host a given model on, billed per second, with a base model server (with NVIDIA L40S GPU and 2 GB of RAM) priced at $1.94 credits per one hour of hosting.

In practice, the Pro subscription is enough to run private models for hours. We only start servers when needed and proactively stop them when inactive to minimize the hosting time.

External Models

We allow to plug & use any external (cloud or private) models by connecting to arbitrary OpenAI-compatible API (which means virtually all major cloud AI providers). This way you have ultimate flexibility to bring in & mix your own plans for OpenAI, Google or other cloud/private models on top of built-in and hosted ones.

External model usage is free of charge.

Tool Pricing

This section describes how credits are used when using tools depending on the tool type. Learn more about tool types on the  Working With Tools page.

Built-In Tools

We provide competitive prices for unified, zero-effort access to a selection of leading cloud tools including Exa and Linkup. This way capabilities such as web search come with a choice of data provider and control over number of results & other parameters.

Credits are spent based on number and type of calls to the tool. Rates per 1K calls are tool-specific and are specified on each built-in tool's page.

In practice, the Starter subscription is enough to cover hundreds of leading tool calls.

Private And Computer Tools

We can host your own private tools from a wide choice of MCP tools (local with stdio transport), letting you tap into the richest & most thriving community of AI tools. We also allow to combine multiple private tools and equip them with arbitrary hardware on Computer tools. We offer very competitive prices while also covering all the hosting-related heavylifting.

Private tool usage is charged by hosting time of the server that we host a given tool on, billed per second, with a base tool server (with 512 MB of RAM) priced at $0.24 credits per one day of hosting. Price of computers with extra RAM is proportional to the amount of RAM picked - for example, one with 1 GB of RAM costs 2X more per hosting duration than base one with 512 MB etc.

In practice, the Pro subscription is enough to run private tool for weeks. We only start servers when needed and proactively stop them when inactive to minimize the hosting time.

Drive Tools

We provide the capability to store files persistently on Drive tools, allowing to plug data storage either directly to AI chats and assistants or as attachment to Computer tools.

Drive tool usage is free of additional charge as long as you're in the Pro or higher plan and your overall data usage - sum of chat uploads and all drives - fits within your plan's storage limit.

Assistant Tools

We allow to turn your assistants into tools. This powerful feature opens a whole world of possibilities by letting you use your refined, specialized & focused assistants in other chats or assistants.

Assistant tool usage is free of additional charge as long as you're in the Pro or higher plan. You are of course charged as usual for model and tool calls made by the actual assistant.

External Tools

We allow to plug & use any external (cloud or private) tools by connecting to arbitrary MCP SSE API. This enables a rich ecosystem of MCP cloud services. This way you have ultimate flexibility to bring in & mix your own plans for various cloud/private tools on top of built-in and hosted ones.

External tool usage is free of charge.