Skip to main content
Glama
Sign In
tru8-mcp

tru8-mcp

by SamYatesSmith
OverviewSchemaSourceRelated ServersDiscussionsAdmin
Python
Remote

tru8-mcp

MCP server for Tru8 Evidence Research — structured evidence research tools for AI agents.

Submit a claim or URL, get back source-traced evidence organized by tier (primary/reporting/commentary) and type (data/official/news/analysis/opinion/academic), with element decomposition and relationship mapping.

Two ways to connect

Hosted (nothing to install). Point any MCP client that supports remote servers at:

https://api.trueight.com/mcp/

Authenticate with your Tru8 API key, sent either as an X-API-Key header or as Authorization: Bearer. Each request is authenticated on its own, so one endpoint serves every user without their keys ever meeting.

Local (this package). Runs on your own machine over stdio, which means your API key never leaves it. Use this if you would rather not send a credential through anyone else's infrastructure.

Both serve the identical tools from the identical code — pick whichever suits.

Related MCP server: Research Oyster

Quick Start

pip install tru8-mcp

Or run without installing:

uvx tru8-mcp

Set your API key:

export TRU8_API_KEY=tru8_sk_...

Create an API key at your Tru8 dashboard under Settings > Developer.

Configuration

Claude Desktop

Add to claude_desktop_config.json:

{
  "mcpServers": {
    "tru8": {
      "command": "tru8-mcp",
      "env": {
        "TRU8_API_KEY": "tru8_sk_..."
      }
    }
  }
}

Cline / Cursor

Add to your MCP settings:

{
  "mcpServers": {
    "tru8": {
      "command": "uvx",
      "args": ["tru8-mcp"],
      "env": {
        "TRU8_API_KEY": "tru8_sk_..."
      }
    }
  }
}

From source

git clone https://github.com/SamYatesSmith/tru8-mcp.git
cd tru8-mcp
pip install -e .
tru8-mcp

Tools

Tool

Description

Typical time

tru8_check

Evidence research for a claim or article URL

15-120s

tru8_get_result

Retrieve completed check with computed analytics

<1s

tru8_get_result_raw

Retrieve raw check data without computed analytics

<1s

A check that is still running after several minutes returns its checkId and status instead of the result. Fetch it later with tru8_get_result (free). Calling tru8_check again with the same claim within ten minutes returns the same check and is not charged twice.

Environment Variables

Variable

Required

Description

TRU8_API_KEY

Yes

API key (tru8_sk_...). Create at dashboard > Settings > Developer.

TRU8_API_URL

No

API base URL. Default: https://api.trueight.com

Security

Store API keys in environment variables or a secrets manager. Never hardcode keys in source code. If a key is compromised, revoke it immediately at your dashboard.

Available Tools

3 tools
tru8_checkA

Evidence research for a factual claim or article URL.

Submit a claim as text or paste an article URL. URLs are auto-detected and trigger article mode: the pipeline extracts claims from the page and auto-selects up to 5 for evidence research.

Tiers (in fallback order):

  • lookup (~£0.02, instant) — cached prior analysis

  • consensus (~£0.03, instant) — cross-user aggregate landscape (k≥3 checks)

  • quick (~£0.07, ~15s) — web search + heuristic classification

  • full (~£0.15, ~60-90s) — 30+ sources, LLM classification, coverage recovery

Charges based on tier actually executed, not tier requested. Set max_tier to control maximum spend per call.

Output structure:

  • claims[].claimMap.elements[] — verifiable sub-claims with state (supported/disputed/unresolved)

  • claims[].claimMap.elements[].evidenceRefs[] — evidence mapped to elements with relationship (supports/challenges/context) and reasoning

  • claims[].evidence[] — sources classified by tier (primary/reporting/ commentary) and type (data/official/news/analysis/opinion/academic)

  • claims[].claimMap.orientation — mechanical summary from element states

  • _meta — execution metadata: executedTier, chargedPence, limitations

Args: claim: A factual claim ("The Earth's average temperature rose 1.1°C since 1880") or an article URL (https://example.com/article). URLs are auto-detected and the pipeline extracts claims from the page content. max_tier: Maximum tier to attempt — "lookup", "consensus", "quick" (default), or "full". max_age_hours: Skip cache hits older than this many hours. If set, lookup hits that are stale will be discarded and the pipeline re-runs at the next tier up to max_tier. compact: If True, strip full evidence arrays from response (smaller payload).

Inputs: claim (required), max_tier, max_age_hours, compact. Returns: result

tru8_get_resultA

Retrieve a previously submitted check with pre-computed analytics.

Returns the full result including a _computed block with tier/type distributions, corroboration groups, diagnostic values, timeline analysis, element state summaries, and per-claim dispositions.

Use this over tru8_get_result_raw when you want structured analytics ready for summarisation or comparison without post-processing.

Args: check_id: UUID returned by tru8_check.

Inputs: check_id (required). Returns: result

tru8_get_result_rawA

Retrieve a previously submitted check without computed analytics.

Returns claims, elements, evidence, and claim maps only. No _computed block. Smaller response payload. Use this when you will compute your own aggregations or only need specific fields from the raw data.

Args: check_id: UUID returned by tru8_check.

Inputs: check_id (required). Returns: result

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    Deploy Server
    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

    TDQS

    A4.6/5.0

    Scored across 3 tools

    Disambiguation4/5

    The three tools have distinct purposes: check submits a claim, get_result returns analytics, and get_result_raw returns raw data. The two retrieval tools are very similar but explicitly differentiated by descriptions, causing minimal ambiguity.

    Naming Consistency5/5

    All tools follow the pattern tru8_<verb>_<modifier> with consistent snake_case. The verbs 'check' and 'get' are clear, and modifiers 'result' and 'result_raw' are descriptive and uniformly applied.

    Tool Count5/5

    Three tools is well-suited for the narrow domain of fact-checking: one for submitting a claim and two for retrieving results in different formats. No unnecessary tools, and the count is sufficient for the core workflow.

    Completeness4/5

    The tool surface covers the primary workflow (submit and retrieve results with or without analytics). However, there is no tool to list or manage previous checks, which is a minor gap for users who need to track multiple checks.

    Maintenance

    ActivityMaintained
    ResponsivenessOutdated data

    Resources

    Related MCP Connectors

    Related MCP Servers

    • Research Oyster

      zandemha2025
      A
      license
      Not graded
      quality
      D
      maintenance
      MCP server for local, source-agnostic research, turning briefs into platform-specific searches and cited evidence dossiers with PostgreSQL storage and optional browser capture.
      MIT

    • llm-sidecar

      awaistechnologist
      A
      license
      Not graded
      quality
      B
      maintenance
      MCP server for grounded, cited AI: answers questions from live web sources, verifies claims, fact-checks documents, searches and reads URLs, summarises, classifies, and extracts fields, with usage tracking and status.
      59 PyPI
      1
      MIT