# FlyMyAI Headless Cloud - platform log

Current version: **1.0.1** (2026-09-25). Newest first; this page shows the latest 10 entries.

- What the platform does today (the current snapshot): https://flymy.ai/skill.md
- Every entry, machine-readable: https://flymy.ai/changelog.json
- Agents: call `whats_new` on the Agents MCP with `since` set to the last version you saw.

## 1.0.1 - 2026-09-25 - Baseline: FlyMyAI Headless Cloud as it runs in production

- **models** - One catalog of generative models: images, video, audio and music, speech, upscaling, embeddings and LLMs. recommend_model picks one for a task, get_pricing shows the price, run_model runs it on your key. Use: `list_media_models`, `recommend_model`, `get_pricing`, `run_model`. [Docs](https://docs.flymy.ai/quickstart)
  Try: "Make a 5-second vertical video of neon koi fish swimming through rain."
- **agents** - Build one agent from an idea: create it, run it and read the run, refine the same agent, then freeze an accepted run into a reusable release and run that release. A frozen agent keeps running in the cloud when your laptop is closed. Use: `create_agent`, `run_agent`, `get_run`, `freeze_agent`, `get_compilation`, `run_frozen`. [Docs](https://docs.flymy.ai/agents/quickstart)
  Try: "Build me an agent that turns a product link into three ad headlines, test it and freeze it."
- **templates** - Open a media template from the gallery at flymy.ai/media (product ads, unboxings, outfit reveals and more), run it as is, clone it and make it your own agent. [Docs](https://docs.flymy.ai/agents/guides/media-templates)
  Try: "Show me the media templates and clone the unboxing one for my sneakers."
- **tools** - Search the connector catalog by what you need done (Slack, Notion, GitHub, Google Workspace, CRMs, payments, web search, browser automation and more), read the exact schema, then run an action directly or inside an agent. Use: `search_tools`, `get_tool_schemas`, `execute_tool`, `add_tool`. [Docs](https://docs.flymy.ai/mcp)
  Try: "Post a summary of my open Linear issues to our team Slack channel."
- **sharing** - Pilot: share one connected account with another person by email from the MCP page in the app. They use only reviewed read actions and pay for their own calls; your credentials never leave FlyMyAI. Sharing with a team is in a private pilot. [Docs](https://docs.flymy.ai/agents/guides/mcp-personal-sharing)
  Try: "Share my Notion connection with alex@example.com for a week."
- **templates** - Share an agent as a template: set_agent_public makes it public, and copy_agent forks an agent into your own account to change as you like. Use: `set_agent_public`, `copy_agent`.
  Try: "Make my headline agent public so a friend can fork it."
- **agents** - Run a frozen agent on a schedule, or call it from your product over the Agents REST API or the Python SDK (package flymyai). Use: `schedule_agent`, `run_frozen`. [Docs](https://docs.flymy.ai/agents/guides/call-from-your-app)
  Try: "Run my headline agent every morning at 9 and post the result to Slack."
- **agents** - An agent can hand parts of a run to subagents and join their results in the same run. subagent_limits on run_agent caps their spend, count and parallelism before anything starts. Use: `run_agent`.
  Try: "Research these five competitors in parallel and give me one comparison table."
- **models** - TypeSafe Jev 1.13 answers typed questions quickly and cheaply: pick one option, give a yes probability or a score. Use it for routing, triage and guardrail checks from an agent or directly. Use: `make_typed_decision`. [Docs](https://docs.flymy.ai/guides/typesafe-jev)
  Try: "Sort these support tickets into billing, technical and sales."
- **tools** - Connect several accounts of one service under aliases, group connections into MCP resource sets and grant a set to agents or to an agent group. Use: `add_tool`, `create_mcp_resource_set`, `replace_mcp_resource_set_members`, `create_agent_group`. [Docs](https://docs.flymy.ai/agents/guides/mcp-resource-sets)
  Try: "Connect my work and personal Gmail and let my inbox agent use only the work one."
- **mcp** - Connect an MCP client to one server, https://mcp-agents.flymy.ai/mcp, with OAuth sign-in or an API key sent as X-API-Key. It carries models, tools and agents; whoami confirms which account is connected. Use: `whoami`, `search_tools`. [Docs](https://docs.flymy.ai/agents/guides/ai-assistants)
  Try: "Which FlyMyAI account am I on, and what is my balance?"
- **mcp** - Bring your own MCP server: register it, connect it and call its tools from chat or from an agent. Use: `add_mcp_server`, `connect_mcp_server`, `call_mcp_server`. [Docs](https://flymy.ai/skill.md)
  Try: "Connect my MCP server at https://example.com/mcp and list its tools."
- **api** - Embed an agent in your product with one builder key: publish a frozen agent as a deployment, identify each of your users by external_user_id, let them connect their own accounts through a hosted link, and run and meter per user. Use: `POST /api/v1/agents/deployments/{deployment_id}/run/`. [Docs](https://flymy.ai/skill.md)
  Try: "Show me the code to call my frozen agent from my Next.js app."
- **compute** - Deploy your own Python model as an auto-scaling GPU endpoint with the FlyMyAI CLI (fma) and pay only for the compute it uses. [Docs](https://docs.flymy.ai/cli/intro)
  Try: "How do I deploy my PyTorch model as a GPU endpoint?"