Skip to content

What’s next for Automedha

More of the work.
Less to operate.

As your work grows, so do the jobs your agents can handle. These are the capabilities coming to Automedha.

Request early access
Where your work can goUpcoming
Upcoming directions for your agent workFollow a link to explore upcoming capabilities for your agent work.Your codeContext & dataCode & webLarger tasksCost controlYour work
More ways to build, run and deliver agent work. Explore a direction to see what’s coming.

Your code

Upcoming

Your own code, on the same runtime.

Not every process fits a catalog agent. Your own agents get the same place to run.

  • Run agents you write.

    Bring your own agent code to managed execution, so custom work doesn’t start another hosting project.

  • Develop with the work in view.

    A local Studio to write, test and inspect your agents before putting them to work.

Context & data

Upcoming

Carry context from one task to the next.

A brief, a working file, and what an agent learned stay useful beyond the task that created them.

  • Remember between tasks.

    Keep useful facts and context in agent memory, so later work builds on what came before.

  • Keep working files organized.

    Manage files beyond a run’s temporary share link, with control over how long they stay available.

  • Work with stored data.

    Read and maintain task data in databases and searchable stores, from structured records to reference material.

Code & web

Upcoming

Give agents the tools a job needs.

Some jobs need a command, a checkout or a browser. Tools that let agents act on those parts of a task.

  • Run the tools the task needs.

    Transform data, execute commands and use dependencies and build tools in isolated sandboxes, including previews of the result.

  • Make changes with the source in view.

    Use coding models with repository context, then read and update private Git repositories as part of the task.

  • Work inside browser applications.

    Use authenticated browser sessions and screenshots for tasks that need to interact with a page.

  • Give more tasks a way to finish.

    Extend tool access with audio, user verification, messaging, repository exports and more ways to manage domains.

Larger tasks

Upcoming

Take on work that needs more than one agent.

A larger assignment needs deeper research, specialist help and a decision from a person. Those steps stay connected.

  • Investigate further and divide the work.

    Explore larger research questions and delegate parts of a task to specialist agents.

  • Bring a person into the decision.

    Review requests for approval or clarification, then let the waiting work continue from its place.

  • Follow the decisions along the way.

    Inspect a richer run history, including who started the work and who stepped in.

Cost control

Upcoming

Pick providers by price. Trace charges to the step.

As more agents do paid work, the choice of tool and the reason for a charge matter to the people operating them.

  • Compare suitable tool providers.

    Route capability requests using prices from more sources, so cost informs which provider does the work.

  • Fund the work together.

    Shared funding and finer cost attribution across team operations, down to the steps that incurred a charge.

A starting point today

Start with a job you already have.

Research briefs, writing, and content packs are in the catalog today. Explore the agents to find an output you can use.

Explore the agents

Early access

Start with the agent you’d run first.

Tell us which task comes first and who the result is for.

Request early accessNo payment details required to request access.

Grounded in Automedha notes

Ask what runs today.

Sources are listed under the answer.