Kalima

Follow a recording as it happens

get_live_transcript

Read-onlyIdempotent

Reads a recording that is running right now, e.g. "what have we agreed so far", "was besprechen wir gerade", "worum ging es in den letzten Minuten". Without sessionId it uses the user's own running recording first; session.ownedByYou false means someone else is recording it. The first call returns the transcript so far (tailSegments starts near the end instead). Passing the returned resumeCursor on the next call returns only what was said since. segments or transcript hold new text; a provisional segment is still being spoken and comes again; revised lists corrections to earlier text by index. session.isLive false means the recording has ended, hasMore true means more is waiting, found false means nothing is being recorded. detail works as on get_transcript and defaults to "speaker".

Input Schema

Name	Required	Description
`limit`	No	How many segments to return (max 300, default 120).
`cursor`	No	nextCursor from the previous call, to get the next page.
`detail`	No	"text": one text block. "speaker": text with a [speaker] label where the speaker changes. "full": segments with index, speaker and startMs.
`sessionId`	No	Recording id, as returned by list_recordings or search_recordings.
`tailSegments`	No	Without a cursor: start this many segments before the end instead of at the beginning, to join a long recording already in progress. Ignored when a cursor is given.

Output Schema

Name	Required	Description	Default
`found`	Yes
`notice`	No
`hasMore`	No
`revised`	No
`session`	No
`segments`	No
`transcript`	No
`resumeCursor`	No

Schema Changelog

Changes observed during successful MCP inspections.

First observedOct 30, 2026

TDQS

A4.6/5.0

Behavior5/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

Annotations already cover read-only/idempotent safety, but the description adds substantial behavioral detail: polling with resumeCursor returns only new text, provisional segments are re-sent, revised entries correct earlier text by index, and session.isLive/hasMore/found flags signal state. This is exactly the context an agent needs to poll correctly.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness4/5

Is the description appropriately sized, front-loaded, and free of redundancy?

Dense but front-loaded: live-reading intent comes first, then response semantics and flags. The three-language example queries are useful for matching user phrasing but make the opening slightly heavier than necessary.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness5/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

Even with an output schema present, the description explains the practically important response fields (segments/transcript, provisional, revised, resumeCursor, session flags) and the default detail level. Nothing an agent needs in order to poll correctly is missing.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters4/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema coverage is 100%, so the baseline is 3, but the description adds meaningful semantics: sessionId fallback to the user's own recording, tailSegments versus cursor interaction, and that detail defaults to 'speaker' and behaves like get_transcript. Minor naming drift ('resumeCursor' versus the schema's 'cursor') keeps it from a 5.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

States a specific verb and resource with scope: 'Reads a recording that is running right now', and the example queries anchor the intent. This cleanly separates it from get_transcript/show_transcript, which handle completed recordings.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines4/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

Explains the no-sessionId default (user's own running recording first) and the tailSegments alternative for joining a long recording already in progress. It does not explicitly name get_transcript as the alternative for finished recordings, though the 'running right now' framing implies it.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

All Kalima tools Inspect

Try in Browser

One endpoint for every MCP toolRayrun runs every MCP server behind one URL and holds the credentials. Each call is decided per client, tool, and argument, then recorded.ray.runAd

Glama MCP Gateway

Add one secure layer between your agents and this server.

Resources

Need Help?