Skip to main content

​
Concept

Trainings are the knowledge base that powers Timely.ai agents. While the system_prompt defines the agent’s behavior and personality, Trainings provide the factual content it uses to respond: company policies, product catalogs, frequently asked questions, support manuals, sales scripts, and any other material the agent needs to consult during a conversation. Each Training is associated with one or more agents and is indexed so the agent can perform semantic searches when it needs specific information — without loading all the content into context with every message.

​
Supported Content Types

You can add trainings in different formats:
  • Free text — any structured or unstructured content (Q&A, policies, instructions)
  • File — PDF, DOCX, TXT, MD uploaded via upload
  • URL — Timely crawls and indexes the page content

​
Lifecycle

create training (processing)
    │
    ▼
semantic indexing
    │
    ▼
ready → agent queries during conversations
    │
    ▼
update content → reprocessing
    │
    ▼
delete (removes from the agent's knowledge base)
The processing status indicates the content is being vectorized. Once it becomes ready, the agent can query it in subsequent conversations.

​
Main Fields

FieldTypeDescription
idstring (UUID)Unique training ID
agent_idstring (UUID)Agent it belongs to
titlestringDescriptive title of the content
typeenumtext | file | url
contentstringText content (for type text)
file_urlstringURL of the indexed file
source_urlstringSource URL (for type url)
statusenumprocessing | ready | error
created_atstringISO 8601 creation timestamp
updated_atstringISO 8601 last update timestamp

​
Available Endpoints

See the left sidebar for the full list of endpoints in this group. Each endpoint has its own page with request/response schema, cURL/JS/Python examples, and an interactive playground.

​
Next Steps

Your first request

If you haven’t made your first call yet, start here.