AgentWorkflow automation / Data & APIs
Perform desktop and document workflow tasks using screen, mouse and keyboard interaction.
- How to use it
- Use the documented agent_s workflow with explicit model and screen settings.
- Pricing
- Unknown total cost. Check hosting, model usage and any service plan before use. A source license does not establish free operation.
- Permissions
- Can control the desktop. Optional local environment executes Python/Bash; review this separately before enabling it.
AgentCoding
Edit a codebase from a terminal conversation, with Git changes and optional lint/test feedback.
- How to use it
- Use the official Aider CLI setup; start it in your project and select a model.
- Pricing
- Total cost is unknown. Check model usage, hosting and any service plan before use. No price, currency or billing unit has been confirmed.
- Permissions
- May read and edit project files, create commits and run configured checks. Hosted models need provider credentials.
AgentCoding
Plan and edit a project in an editor or terminal, with file changes and command execution.
- How to use it
- Choose an official Cline client; the SDK is a separate developer resource.
- Pricing
- Total cost is unknown. Check model usage, hosting and any service plan before use. No price, currency or billing unit has been confirmed.
- Permissions
- File edits and terminal commands require approval by default in the documented Plan/Act workflow; auto-approval changes that boundary. Model access and workspace permissions need review.
AgentResearch / Workflow automation
Research a topic and work on extended tasks through a configured agent workspace.
- How to use it
- Use the runnable DeerFlow application and setup wizard; this is not a listing for its embedding APIs.
- Pricing
- Total cost is unknown. Check model usage, hosting and any service plan before use. No price, currency or billing unit has been confirmed.
- Permissions
- Setup selects sandbox mode, bash access and file-writing tools. Model/search keys and any gateway or tool access must be configured deliberately.
AgentFinance & Crypto / Research / Data & APIs
Run the finrobot_equity workflow to fetch company financial data, generate model-written analysis and produce HTML/PDF equity research reports.
- How to use it
- Specific finrobot_equity command-line report pipeline; FMP and OpenAI API keys are documented requirements. This entry is not the entire FinRobot framework or a hosted account.
- Pricing
- Total usage cost is unknown; model, data-provider and compute charges can apply. Open-source licensing does not mean free operation.
- Permissions
- The CLI report workflow calls financial/model services and writes local reports; optional sentiment requests are external. If the documented web/deploy.sh path is chosen, it additionally installs dependencies, opens a web listener, creates a user database (documented password hashes/GitHub OAuth) and writes request logs. These web effects are conditional, not required for every CLI report. No brokerage execution is established.
AgentCoding
Find relevant files, change code and run project checks from a terminal request.
- How to use it
- Use Freebuff CLI; the underlying Codebuff SDK is a separate resource.
- Pricing
- The README describes ad-supported included models without a subscription or API key. Limits, region and model terms vary; paid account/model balances and total costs are not verified.
- Permissions
- Can read code, edit files and run commands. The README describes service processing of prompts/code and ad-related data use; inspect the linked privacy terms before sending private work.
AgentCoding / Research
Explore and edit a codebase with a terminal coding agent and selected tools.
- How to use it
- Use the official CLI setup and its supported account or API authentication.
- Pricing
- The README describes a personal-account free tier. Eligibility, current limits, paid models and total costs are not independently confirmed.
- Permissions
- Workspace tools may read files, edit code and execute commands; model authentication is required.
AgentCoding / Research / Data & APIs / Workflow automation
Work on coding, research or data tasks using a local agent and chosen extensions.
- How to use it
- Follow the official desktop or CLI quickstart and choose the extensions you need.
- Pricing
- Total cost is unknown. Check model usage, hosting and any service plan before use. No price, currency or billing unit has been confirmed.
- Permissions
- Provider credentials or a supported account connection may be needed. Extension permissions and the exact approval defaults still need checking.
AgentResearch
Research a question and create a report with links to its sources.
- How to use it
- Choose web or Python access in the official README. Default examples configure a model provider and Tavily search credentials.
- Pricing
- Pricing not confirmed. Model, search API and hosting charges depend on the services and usage.
- Permissions
- Based on the documented research workflow, queries are sent to configured model and search services and information websites are accessed. Local-document mode reads selected material; check what data is sent to each service.
AgentWorkflow automation / Research / Coding
Handle research and workflow tasks with a local conversational agent and selected tools.
- How to use it
- Use the official interactive CLI and configure the chosen provider and tools.
- Pricing
- Unknown total cost. Check hosting, model usage and any service plan before use. A source license does not establish free operation.
- Permissions
- Can use terminal tools and persistent memory. Optional messaging and schedules require separate credentials and approval.
AgentResearch / Workflow automation
Answer research questions using web or selected documents, with optional recurring research.
- How to use it
- Use the official web app or follow its self-hosting guide.
- Pricing
- Total cost is unknown. Check model usage, hosting and any service plan before use. No price, currency or billing unit has been confirmed.
- Permissions
- May process connected documents and send questions to chosen model/search services. Exact permissions and delivery integrations need checking.
AgentCoding / Workflow automation
Work on coding tasks with an interactive agent that keeps memory across sessions.
- How to use it
- Start the documented Letta CLI in a project and configure its model connection.
- Pricing
- Unknown total cost. Check hosting, model usage and any service plan before use. A source license does not establish free operation.
- Permissions
- Memory persists; hooks, tools and permission modes can enable external actions. Review credentials and retained data.
AgentCoding
A lightweight coding agent you can use from the command line.
- How to use it
- Follow the official quickstart, then use the mini CLI or Python interface with a configured model provider.
- Pricing
- Pricing not confirmed. Model calls and computing costs are charged separately.
- Permissions
- Actions use shell commands or subprocesses. Based on those documented capabilities, access depends on the chosen local or sandbox environment and may include reading and writing files, running commands and network access.
AgentWorkflow automation
Handle conversational tasks and scheduled workflows through a self-hosted agent.
- How to use it
- Use the bundled WebUI or terminal agent, not the standalone Python SDK.
- Pricing
- Total cost is unknown. Check model usage, hosting and any service plan before use. No price, currency or billing unit has been confirmed.
- Permissions
- Tools can access files, shell, web and chat channels. Choose workspace/access mode and review channel credentials before enabling them.
AgentCoding
Work on coding tasks in a terminal using a selected model and agent harness.
- How to use it
- Start the current Rust CLI with interpreter; follow its provider and sandbox guides.
- Pricing
- Total cost is unknown. Check model usage, hosting and any service plan before use. No price, currency or billing unit has been confirmed.
- Permissions
- The README links permission and sandbox configuration. Exact defaults and provider credential needs remain to be checked before allowing edits or commands.
AgentCoding
Explore a codebase or make changes through a terminal coding agent.
- How to use it
- Choose Plan for analysis or Build for development in the official CLI.
- Pricing
- Total cost is unknown. Check model usage, hosting and any service plan before use. No price, currency or billing unit has been confirmed.
- Permissions
- Build is described as full access. Plan disables file edits by default and asks permission for bash commands. Review model credentials and workspace access separately.
AgentCoding
Work on coding tasks through the Agent Canvas interface.
- How to use it
- Use the official Agent Canvas guide to access the OpenHands Agent and configure a model. The SDK library itself is a separate resource.
- Pricing
- Pricing not confirmed. Model, infrastructure and optional cloud service charges need checking separately.
- Permissions
- Native setup can access the host file system; Docker setup can access mounted projects. Model connections need credentials. Terminal and file-editing tools can change the workspace.
AgentResearch / Workflow automation
Attempt research and browser workflow tasks from a terminal goal.
- How to use it
- Follow the single-agent python main.py quickstart.
- Pricing
- Unknown total cost. Check hosting, model usage and any service plan before use. A source license does not establish free operation.
- Permissions
- Model APIs require credentials; selected tools may access the browser, network and local files.
AgentCoding
Work on coding tasks and project changes through a terminal agent.
- How to use it
- Use the official terminal setup and configure a supported model connection.
- Pricing
- Unknown total cost. Check hosting, model usage and any service plan before use. A source license does not establish free operation.
- Permissions
- Can operate on workspace files and tools. Credentials and approval defaults depend on configuration.
AgentWorkflow automation / Data & APIs
Carry out browser tasks and extract information using natural-language workflow goals.
- How to use it
- Use the runnable UI and browser task workflow; the Playwright-compatible SDK is a separate interface.
- Pricing
- Total cost is unknown. Check model usage, hosting and any service plan before use. No price, currency or billing unit has been confirmed.
- Permissions
- May use website sessions or stored credentials, submit forms and download files. Review each workflow and target-site permission before use.
AgentResearch
Research a topic, produce an outline and generate a cited article.
- How to use it
- Run the documented STORM topic workflow or demo; the knowledge-storm package alone is not the listed experience.
- Pricing
- Total cost is unknown. Check model usage, hosting and any service plan before use. No price, currency or billing unit has been confirmed.
- Permissions
- Research sends queries to model/search services and writes article artifacts. Example provider keys are required for those services; local permissions need review.
AgentCoding
Attempt to fix a repository issue using a language model and coding tools.
- How to use it
- Use the original SWE-agent command-line issue workflow.
- Pricing
- Total cost is unknown. Check model usage, hosting and any service plan before use. No price, currency or billing unit has been confirmed.
- Permissions
- Needs access to the chosen repository and tool runtime; model authentication and exact sandbox/approval settings require checking.
AgentFinance & Crypto / Research / Data & APIs
Run the TradingAgents CLI research workflow to combine market, fundamental and news analysis into a model-generated research/decision output for a ticker and date.
- How to use it
- Local CLI plus model-provider access. Default stock/fundamental/news vendor is yfinance; optional Alpha Vantage and configured macro/prediction-market sources have separate requirements.
- Pricing
- Total usage cost is unknown; model, data-provider and compute charges can apply. Open-source licensing does not mean free operation.
- Permissions
- Calls configured data/model services and persists local decision logs/cache; optional checkpoints write SQLite. A live broker/account connection is not established by this listing.
AgentWorkflow automation / Data & APIs
Automate document and application tasks on a Windows desktop.
- How to use it
- Use the README’s standalone UFO² setup and python -m ufo task workflow.
- Pricing
- Unknown total cost. Check hosting, model usage and any service plan before use. A source license does not establish free operation.
- Permissions
- Desktop automation can read and change application content; model credentials and screen access need review.
AgentWorkflow automation / Testing
Use a visual desktop agent for computer and browser automation tasks.
- How to use it
- Use the UI-TARS Desktop quickstart for local or remote computer/browser operators.
- Pricing
- Unknown total cost. Check hosting, model usage and any service plan before use. A source license does not establish free operation.
- Permissions
- Screen contents and actions can reach the model and affect applications. Operator credentials and permissions need review.