---
title: "Vectara MCP Server: review, setup & limitations · Undominated.ai"
canonical: https://undominated.ai/mcp-servers/vectara-mcp/
description: "Runs Vectara RAG queries and semantic search over named corpora, and calls Vectara's hallucination-correction and factual-consistency APIs, from an MCP client."
---

# Vectara MCP Server: review, setup & limitations · Undominated.ai

> Runs Vectara RAG queries and semantic search over named corpora, and calls Vectara's hallucination-correction and factual-consistency APIs, from an MCP client.

[← Explore all mcp servers](/mcp-servers/)

RETRIEVAL AND GROUNDING / Vectara

# Vectara MCP Server

Runs Vectara RAG queries and semantic search over named corpora, and calls Vectara's hallucination-correction and factual-consistency APIs, from an MCP client.

 [See setup guidance ↓](#setup)[Original source ↗](https://github.com/vectara/vectara-mcp)

SOURCE REVIEW

 Reviewed 2026-10-13
 Evidence 3 linked sources
 Publisher Vectara
 Licence [Apache-2.0 ↗](https://github.com/vectara/vectara-mcp/blob/2b5f109e1fca388dea0931f32749e575a7cc2c1d/LICENSE)
 Revision 2b5f109e1fca
 [Read what was—and wasn’t—checked ↓](#review)

“Vectara-MCP provides any agentic application with access to fast, reliable RAG with reduced hallucination, powered by Vectara's Trusted RAG platform, through the MCP protocol.”

 [Vectara · upstream description ↗](https://github.com/vectara/vectara-mcp/blob/2b5f109e1fca388dea0931f32749e575a7cc2c1d/README.md) Our analysis follows below.

01 / THE REASONING

## Why this made the selection.

 - Published in Vectara's own repository under Apache-2.0; the README at this commit lists six tools, including correct_hallucinations and eval_factual_consistency, which check generated text against supplied source documents.
- The README states that HTTP transport is the default with bearer-token authentication on and a 100 requests/minute default rate limit, and that stdio must be enabled explicitly with --stdio.

### A good fit for

 - Grounding an agent's answers in Vectara corpora and checking generated text against its source documents
- Running semantic search over Vectara corpora without generation, using search_vectara

### Weigh up before choosing

 - setup_vectara_api_key takes the API key as a tool argument, which puts it in the conversation; the VECTARA_API_KEY environment variable avoids that.
- The README names the default generation preset for ask_vectara as vectara-summary-table-md-query-ext-jan-2025-gpt-4o, so a generation call uses the model behind that preset unless overridden.
- The server was not installed, no Vectara call was made, and no hallucination or consistency result was checked.

02 / THE REVIEW RECORD

## What we actually inspected.

Source review has boundaries. A clear record is more useful than a “safe” badge.

### Material inspected

 - README.md at 2b5f109e1fca388dea0931f32749e575a7cc2c1d
- LICENSE at the same commit
- pyproject.toml at the same commit

### Our findings

 - python -m vectara_mcp starts HTTP on 127.0.0.1:8000 with authentication enabled; --no-auth disables it and the README marks that development only.
- VECTARA_ALLOWED_ORIGINS configures CORS and VECTARA_AUTHORIZED_TOKENS adds accepted bearer tokens for the HTTP transport.
- pyproject.toml names vectara-mcp 0.2.1, Python >=3.11; it has no license field, and the LICENSE file is Apache-2.0.

### Not established by this review

 - pip install
- Any Vectara query
- The hallucination-correction and factual-consistency tools

The review applies to the material and revision named here. A newer upstream release can change its behavior.

03 / PUT IT TO WORK

## Connect a server deliberately.

[Upstream setup instructions ↗](https://github.com/vectara/vectara-mcp/blob/2b5f109e1fca388dea0931f32749e575a7cc2c1d/README.md)

DOCUMENTED COMMAND

 pip install vectara-mcp Copy command ↗

Copying does not execute this command. It may retrieve a newer version than the reviewed source.

 - Install the package with the README command above. It was not run.
- For Claude Desktop, add the README's stdio block with VECTARA_API_KEY in env. For HTTP, keep authentication on and set VECTARA_ALLOWED_ORIGINS. Do not commit the key.

### Before you start

 - Python 3.11 or later
- A Vectara API key and corpus keys

### Documented configuration example

Use the client and authentication context described in the setup guidance. Configuration formats vary between hosts.

 {
 "mcpServers": {
 "Vectara": {
 "command": "python",
 "args": [
 "-m",
 "vectara_mcp",
 "--stdio"
 ],
 "env": {
 "VECTARA_API_KEY": "your-api-key"
 }
 }
 }
} Copy configuration ↗

### Compatibility

Claude Desktop · MCP clients over HTTP, SSE or stdio

### Implementation

Python

### Transports

HTTP · SSE · stdio

### Local process with a Vectara API key

 - Queries the Vectara corpora named in each call with the configured key
- Sends generated text and source documents to Vectara's hallucination-correction and factual-consistency APIs

### Cost model

Apache-2.0 server source. Vectara account usage applies; no rate was reviewed.

04 / FOLLOW THE EVIDENCE

## The source trail.

Our notes are separate from the original resource. Check upstream before adopting a new version.

 - [Vectara MCP README ↗](https://github.com/vectara/vectara-mcp/blob/2b5f109e1fca388dea0931f32749e575a7cc2c1d/README.md) Checked 2026-10-13 https://github.com/vectara/vectara-mcp/blob/2b5f109e1fca388dea0931f32749e575a7cc2c1d/README.md Supports: Tools, Transport defaults and authentication, Claude Desktop configuration, Default generation preset
- [Vectara MCP LICENSE ↗](https://github.com/vectara/vectara-mcp/blob/2b5f109e1fca388dea0931f32749e575a7cc2c1d/LICENSE) Checked 2026-10-13 https://github.com/vectara/vectara-mcp/blob/2b5f109e1fca388dea0931f32749e575a7cc2c1d/LICENSE Supports: Apache License 2.0
- [Vectara MCP pyproject.toml ↗](https://github.com/vectara/vectara-mcp/blob/2b5f109e1fca388dea0931f32749e575a7cc2c1d/pyproject.toml) Checked 2026-10-13 https://github.com/vectara/vectara-mcp/blob/2b5f109e1fca388dea0931f32749e575a7cc2c1d/pyproject.toml Supports: Package vectara-mcp 0.2.1, Python >=3.11

KEEP COMPARING

## Other approaches to consider.

Related by category or shared topics. These are alternatives to inspect, not a measured quality order.

 [### Pinecone Assistant MCP Server ↗ Retrieves document snippets from a named Pinecone Assistant for an MCP client through one tool, run as a locally built Docker image or Rust binary.](/mcp-servers/pinecone-assistant-mcp/)[### Amazon Bedrock Knowledge Base Retrieval MCP Server ↗ Lists and queries Amazon Bedrock Knowledge Bases that are tagged for MCP use, with data-source filtering, optional reranking and agentic retrieval on managed knowledge bases.](/mcp-servers/awslabs-bedrock-kb-retrieval/)[### Context7 MCP ↗ Resolves a library identity and retrieves documentation relevant to a specific coding question through Context7's hosted index.](/mcp-servers/context7/)

 [AI Tools ↗](/tools/)[Skills ↗](/skills/)[Agents ↗](/agents/)[MCP Servers ↗](/mcp-servers/)[Workflows ↗](/workflows/)

## Continue your investigation

 - [Inspect resource evidence](/tools/)
- [Choose a host](/agents/)
- [Inspect reusable workflows](/skills/)
- [Use Undominated data](/api/)
