This listing identifies a third-party MCP server. MCPBundles is not affiliated with, sponsored by, or endorsed by the service or its operator unless stated on this page.
This server may ask you to sign in when your AI client connects.
Last probed Nov 3, 2026 · mcp.dapta.dev
140tools discovered
For AI agents — MCPBundles Directory API
For MCP client apps (human connect):
MCP connect URL: https://mcp.dapta.dev/mcp
Skill page: https://www.mcpbundles.com/skills/dapta-mcp-9e91db5390
Explore capabilities (140)
read 18
write 1
other 6
Showing 25 of 140
Showing 25 of 140 tools from the public preview. Connect the server to inspect the full live tool list.
Get10
List7
Temporal2
Other capabilities6
Create Voice Agent
write
create_voice_agent
Create a new voice agent in the caller's Dapta workspace, fully configured.
Part of the create-voice-agent workflow. If you haven't already, call
get_guidance("create-voice-agent") first — it covers template choice,
company-info gathering, and the confirm step.
WARNING: This creates a real, billable voice agent and there is NO
idempotency — every successful call creates a NEW one. Confirm with the user
before calling, and never retry on your own judgement: read `retryable` on
an error and do w
Description may be truncated on the public preview.
Create Voice Agent — Create a new voice agent in the caller's Dapta workspace, fully configured.
Part of the create-voice-agent workflow. If you haven't already, call
get_guidance("create-voice-agent") first — it covers template choice,
company-info gathering, and the confirm step.
WARNING: This creates a real, billable voice agent and there is NO
idempotency — every successful call creates a NEW one. Confirm with the user
before calling, and never retry on your own judgement: read `retryable` on
an error and do w
Preview Update Voice Agent — Preview a voice-agent update WITHOUT committing.
Returns a server-truth diff (current vs proposed) and a single-use preview_token
(signed, valid for a limited window). Tokens cannot be forged — only the
server can issue valid ones.
PROTOCOL (mandatory for the calling skill):
1. Call this tool with the proposed changes.
2. Show the user the returned `diff`.
3. Ask for explicit confirmation in chat.
4. On user yes -> call commit_update_voice_agent(preview_token=...).
5. On user no / pr
Commit Update Voice Agent — Commit a voice-agent update using a preview_token from preview_update_voice_agent.
The token must be unexpired and from the same workspace as the preview.
THIS TOOL RE-READS THE AGENT AFTER WRITING IT, and `success` means "what is
live is what you asked for" — not "the platform answered 200". The platform
has answered `{"success": true, "item": null}` for a write that never
happened (a whole event nearly ran with no post-call extraction), so the
only thing that settles a write is reading it ba
List Phone Numbers — Every phone number this workspace holds, and who answers each one.
Answers "who picks up if I call this number?" and "which number can this
agent call from?" without opening the dashboard. One entry per physical
line, even when the same line lives in several tables.
A workspace holds up to three KINDS of number and each one is assigned in a
different place — the single most useful thing in this output:
- ``purchased`` — bought inside Dapta. Routes inbound on the number
itself. Assigned
List Brains — Knowledge bases (brains) in this workspace and the state of their sources.
Each brain reports how many sources are complete, processing, pending or
failed, plus a one-word ``health``:
- ``complete`` every source loaded.
- ``partial`` finished, but some sources failed.
- ``in_progress`` still loading, and something is actually working on it.
- ``stalled`` sources queued with nobody working on them and nothing
new for 20 minutes. That is a platform-side stall, not a slow c
Get your MCP into directories
A working endpoint is step one. Directory coverage is the coordinated launch across ChatGPT, Claude, Cursor, the MCP Registry, and community indexes.
Dapta is an MCP server at mcp.dapta.dev. MCPBundles discovered 140 tools on the live endpoint.
How do I connect Dapta to my AI agent?
Use the MCP endpoint listed on this page in your MCP client configuration. One-click install pills support Claude, Cursor, VS Code, and other hosts. Copy the remote MCP URL if your client needs a manual entry.
How many tools does Dapta provide?
Operate this server?
Request removal if you do not want Dapta listed here, or claim the listing if you want to manage it on MCPBundles.
Operate Dapta?
This server appears in the MCPBundles directory. Verify you operate it to take over the listing — name, description, logo, contact email, and skill content. We email a 6-digit code to a maintainer address your server publishes in /.well-known/security.txt or /.well-known/mcpbundles.json. Free, takes about a minute.
MCPBundles publishes this directory for MCP discovery. Except where we host or operate an offering, third-party services run under their own terms. Product and company names on this page are used in a descriptive, identifying way (including under nominative fair use where applicable); they remain the property of their owners. Nothing here grants you rights in those marks, and nothing here is an offer to sell a third party's services. Terms · Request removal
More in developer tools
Other MCP servers in this category from the directory index
Get Brain — One knowledge base, with the evidence that tells a stall from slowness.
Returns the brain, its source counts, the agents using it and a
``diagnosis``: how many sources a worker is actually holding
(``with_lease``), the ``retry_count`` histogram, the distinct
``error_message`` values with how many sources each one hit, how many
sources are vectorized, and a one-sentence ``verdict`` written to be pasted
into a support report.
The distinction this exists for: sources in ``processing`` with
``retr
List Whatsapp Templates — WhatsApp templates available to this account, with their Meta status.
Outside a 24-hour window a template is the only way to open a WhatsApp
conversation, so "which one can I actually send?" is the question. Each row
carries ``status`` (Meta's verdict), ``dapta_status`` and the ``sendable``
flag that is true only when both are green.
Templates are stored per ACCOUNT, so this is every template the workspaces
under this account share — not one workspace's private list.
Each row also carries wha
List Agents — List all voice agents in the caller's workspace.
Optional ``search_term`` filters by name or agent UUID
(case-insensitive substring match against the name field). Returns
``count`` (returned), ``total`` (workspace-wide), and ``is_truncated``
when ``count < total``.
Every row carries ``created_by`` and ``updated_by`` as ``{name, email}``,
read from the agent's version history, plus ``created_by_at`` /
``updated_by_at``. USE THEM IN A SHARED WORKSPACE: several Dapta workspaces
hold agents belong
Get Agent — Get full details for a single voice agent in the caller's workspace.
By default returns the agent config WITHOUT the ``instructions`` prompt
(which can be 5-100 KB). Pass ``include_prompt=True`` to include it —
typically only when the user explicitly asks "show me the prompt".
``post_call_analysis_data`` is slimmed (name + type + choices only) unless
``full_schema=True``, which returns it whole (description, examples,
instruction included).
Pass ``include_llm=True`` for the HALF OF THE CONFIG
Get Agent Post Call Schema — Get the post-call analysis schema (extraction variables) for an agent.
Get Agent Complete Analysis — Get a complete lightweight analysis for an agent.
Combines profile + extraction-variable list + call statistics + outcome
distribution. Does NOT include the instructions prompt (use ``get_agent``
with ``include_prompt=true`` for that).
List Calls — List calls in the caller's workspace, filtered, paged and exportable.
Filter in the QUERY, not by eye. Every parameter below is a WHERE clause,
so "voicemails longer than 30 seconds last week for these two agents" is
one call rather than 500 rows you read through.
Date strings accept ``YYYY-MM-DD`` or full ISO timestamps, and the day
boundaries are inclusive on both ends. Use ``include_test_calls=true`` for
agent debugging; the default (false) is what business KPIs want — but note
it only drop
Export Calls — Export many calls at once as CSV or JSONL text.
Same filters as ``list_calls`` — see that tool for what each one means.
This is the one to reach for when the question is "give me every call for
these agents since we went live so I can grade them one by one", which
otherwise means twenty hand-paged requests with no guarantee they did not
overlap.
``format`` is ``"jsonl"`` (default) or ``"csv"``. ``fields`` defaults to
``"summary"`` and also takes an explicit column list; ``"full"`` is
refused,
Get Call — Get ONE call by id, with its transcript, analysis and tool calls.
Use this whenever you already have a call id — from a campaign log, from a
calls-history link, from a customer complaint. Listing calls and filtering
by eye is the wrong tool for that question.
``call_id`` accepts any of the three forms the same call is referred to:
the bare UUID (what the calls-history URL shows), the prefixed
``call_<uuid>`` that campaign logs carry, or the Dapta call id. You do not
need to know which one you
Search Calls By Phone — Search calls by phone number (matches both ``from_number`` and ``to_number``).
This is the contact-history read: every call to or from one number. Strips
non-digits from ``phone_number`` and does a substring LIKE match, so the
last 10 digits are enough — and a number stored as ``+521...`` still matches
the same line written ``+52...``.
``include_detail=false`` (default) returns a compact row per call: ids,
numbers, direction, status, disconnection reason, duration, credits,
success and voicema
Get Agent Call Statistics — Get aggregated call statistics by agent (admin-panel parity).
Returns one row per agent with totals, connection/success rates, contact
counts, average duration, and credits. Pass ``agent_ids`` to filter to a
subset; omit for all agents in the workspace.
List Calls With Audio — List calls that have an audio recording, with transcript previews.
List Call Examples By Category — List example calls where a post-call variable equals a specific value.
Used to show concrete examples after viewing the aggregate distribution
from ``aggregate_variable_results``.
Get Agent Call Volume Check — Check if an agent has enough call data for temporal analysis.
Returns total_calls, span_days, distinct_hours, distinct_days, and a
qualitative ``analysis_recommendation`` (full / partial / insufficient).
Pass the agent's Dapta UUID as ``agent_id`` (the server also accepts the
agent's platform id).
Get Hourly Call Distribution — Get hourly call distribution for an agent (timezone-aware).
Returns one row per local_hour (0-23) with totals, connected/successful
counts, and rate percentages. Use after ``get_agent_call_volume_check``
confirms enough data is available.
Get Weekday Call Distribution — Get weekday call distribution for the workspace (timezone-aware).
Returns one row per ISO weekday (1-7) with totals, connected/successful
counts, rate percentages, and average call duration.
Temporal Heatmap — Get a weekday x hour call-density heatmap for the workspace (timezone-aware).
Temporal Hourly By Org — Get the workspace-level hourly call distribution (timezone-aware).
Get Call Analytics — Combined call-analytics report — one call instead of several (Phase 3).
Part of the voice-call-analytics workflow. Returns the connection-rate
breakdown, call outcomes, and workspace satisfaction together, so you do not
need to chain get_connection_rate + get_call_outcomes + workspace_satisfaction
for a standard "how are my calls doing?" question. Pass agent_id to scope the
connection-rate + outcomes to one agent (satisfaction is workspace-wide).
Discover Variables From Calls — Discover output variables by analyzing actual call extractions.
Samples up to ``sample_size`` calls and tabulates which custom_analysis_data
keys appear, their inferred type, and up to 10 sample values per variable.
Use to understand what variables a deployed agent actually emits — vs.
what its formal post_call_analysis_data schema declares.
Showing the first 25 tools — full list and schemas: Directory API.
MCPBundles probed 140 tools on the live server. The tool list on this page reflects what was discovered at the last refresh — connect your client to see the full set available to your session.
What authentication does Dapta require?
Dapta may require signing in to the provider before tools can run. Connect through MCPBundles or your MCP client and complete any provider login when prompted.
Don't want this listed?
Request removal from the MCPBundles directory. We stop publishing this listing on mcpbundles.com.
MCPBundles is an independent platform built on the open Model Context Protocol standard. MCPBundles is not affiliated with, sponsored by, or endorsed by Anthropic PBC.