generated: '2026-08-30'
method: searched
source: https://help.tellius.com/api
name: Tellius rate limits
limit_count: 0
note: 'Tellius publishes NO API rate limits. Searched the full 565 KB documentation
  index (help.tellius.com/llms.txt) for "rate limit", "throttl", "quota per", "429"
  and "Retry-After" — zero hits — and read every page under /api. No limits section,
  no per-key or per-account ceiling, and no exhaustion status code is documented. This
  is an honest zero, recorded rather than omitted.'

limits: []

response_headers:
  documented: []
  observed: []
  observed_note: 'Probed a Tellius deployment host unauthenticated on 2026-08-30. The
    401 responses carry no X-RateLimit-*, no RateLimit-* and no Retry-After. Only
    standard headers plus strict-transport-security, an etag and a GCLB session
    cookie were returned, so there is no runtime rate-limit signal for an agent to
    read.'
exhaustion_status: null

capacity_limits_instead:
  note: 'What Tellius bounds is CAPACITY per tier, not request rate. These are the
    ceilings an integrator actually meets, published on the pricing page rather than
    in the API reference.'
  limits:
  - scope: per-account (Premium tier)
    dimension: concurrent AutoML jobs
    limit: 5
    source: https://www.tellius.com/pricing
  - scope: per-account (Premium tier)
    dimension: concurrent data pipeline jobs
    limit: 5
    source: https://www.tellius.com/pricing
  - scope: per-account (Premium tier)
    dimension: live/pushdown query rows in live-mode
    limit: 50000000
    source: https://www.tellius.com/pricing
  - scope: per-account (Premium tier)
    dimension: in-memory cache storage
    limit: 10GB
    source: https://www.tellius.com/pricing
  - scope: per-account (Premium tier)
    dimension: users
    limit: 10
    source: https://www.tellius.com/pricing
  - scope: per-request
    dimension: ML prediction file upload size
    limit: 250MB
    source: https://help.tellius.com/api/ml-model-apis
  - scope: per-request
    dimension: 'ML prediction result page size (nrows)'
    limit: 'default 200, no maximum documented'
    source: https://help.tellius.com/api/ml-model-apis

soft_guidance:
- guidance: 'Avoid firing many deep insights at once. Long-running calls hold resources
    for their whole duration. Queue them rather than launching a large batch in
    parallel, or fast queries issued at the same time may be left waiting.'
  applies_to: tellius_deep_insight (MCP)
  source: https://help.tellius.com/kaiya/tellius-mcp-server
  note: 'This is the only concurrency guidance Tellius gives an agent, and it is prose
    advice with no enforced number attached.'
- guidance: 'Set a generous client timeout — a timeout tuned for a simple query will
    cut off a deep insight before it finishes.'
  applies_to: tellius_deep_insight (MCP)
  source: https://help.tellius.com/kaiya/tellius-mcp-server

findings:
- 'An agent calling Tellius cannot discover its budget: there is no documented limit,
  no header, and no 429 contract. Because Tellius is customer-deployed, effective
  limits are whatever the customer''s cluster and tier capacity allow — but that is
  inferred, not stated.'