Skip to main content
Glama
ChrBoebel

Optical Context MCP

by ChrBoebel
OverviewSchemaSourceRelated ServersDiscussionsAdmin
Python
Local

Optical Context MCP is built for one specific job: turning large, visually structured PDFs into a smaller set of retrievable packed images for agent workflows.

It reads a local PDF, runs OCR with Mistral, recomposes the extracted text and figures into dense PNGs, and exposes those artifacts over MCP for batch retrieval.

What It Does

  • reads a local PDF from the MCP host machine

  • extracts page markdown and embedded images with Mistral OCR

  • packs that content into dense PNGs that preserve visual grouping

  • optionally sizes embedded figures with a bundled technical-document model

  • stores a manifest and temp job artifacts for follow-up retrieval

  • lets an agent pull only the packed images it needs

Related MCP server: VisionSieve MCP

Where It Fits

Use it for:

  • operating manuals

  • scanned handbooks

  • product catalogs

  • PDF slide decks

  • visually structured OCR-heavy documents

Skip it for:

  • tiny PDFs

  • clean text-native PDFs where normal extraction is enough

  • workflows that require exact page-faithful rendering

  • cases where OCR cost is not justified

Example Result

The image below shows a real local validation run on a public research paper with dense text, figures, charts, and page-level visual structure. The packed image on the right consolidates the seven source pages shown on the left.

Example local run facts from the generated manifest:

  • source paper pages: 22

  • previewed source page range: 15 to 21

  • extracted images: 30

  • packed output images: 6

  • example packed image size: 986x1084

  • example packed image file size: 536,697 bytes

This example shows the intended workflow: take a long, visually structured PDF and compress it into a smaller set of retrievable packed images that still preserve the visual structure of the source.

Install

python -m pip install optical-context-mcp

Install with the adaptive sizing runtime:

python -m pip install "optical-context-mcp[ml]"

Run without installing:

uvx optical-context-mcp

  • MISTRAL_API_KEY is required for compress_pdf

  • packed images are always stored locally under the system temp directory

  • compress_pdf returns up to 30 packed images inline by default

  • the adaptive sizing checkpoint is bundled with the package

  • adaptive sizing activates automatically when torch and torchvision are available

  • set OPTICAL_CONTEXT_DISABLE_ADAPTIVE_SIZING=1 to force the legacy fixed sizing

  • set OPTICAL_CONTEXT_ADAPTIVE_MODEL_PATH=/path/to/model.pt to override the bundled checkpoint

For pinned shared setups:

uvx --from optical-context-mcp==0.1.5 optical-context-mcp

Run

Default transport is stdio:

optical-context-mcp

Claude Code

Register the server in a project:

claude mcp add -s project optical-context -- uvx optical-context-mcp

Typical use:

  1. call compress_pdf

  2. inspect the returned manifest

  3. fetch packed images with get_packed_images

MCP Tools

  • compress_pdf: run OCR plus recomposition and create a stored job

  • get_job_manifest: load metadata for an existing job

  • get_packed_images: fetch one or more packed PNGs from an existing job

How It Works

flowchart LR
    A["Local PDF"] --> B["Mistral OCR"]
    B --> C["Page markdown + embedded images"]
    C --> D["Recomposition engine"]
    D --> E["Dense packed PNG images"]
    E --> F["Stored job artifacts"]
    F --> G["Agent fetches manifest or image batches over MCP"]

Why Packed Images Instead Of Just OCR Text

  • section grouping

  • table-like layout

  • captions near figures

  • visual adjacency between text and embedded graphics

For many vision-capable agents, that is a better intermediate format than a plain OCR dump.

Current Scope

  • depends on Mistral OCR

  • currently handles local file paths, not remote uploads

  • stores artifacts in the local system temp directory by default

  • optimized for compression and retrieval, not final polished markdown generation

  • quality depends on OCR quality and the visual density of the source document

  • adaptive sizing falls back safely to fixed medium image sizing when the ML runtime is absent

Roadmap

  • make the OCR layer provider-agnostic so different OCR backends can be swapped behind the same MCP workflow

Feedback

Tried Optical Context in a document workflow? You can voluntarily share what worked and what got in the way using the feedback form. Include your package version, MCP client and a brief description of the document type. Do not attach private PDFs or API keys. Feedback is public on GitHub.

See the changelog for changes between published versions.

Development

uv venv --python /opt/homebrew/bin/python3.11 .venv
uv pip install --python .venv/bin/python -e ".[dev]"
.venv/bin/python -m pytest

Publishing

Keep pyproject.toml and both versions in server.json synchronized. Run the Publish PyPI workflow on the reviewed release commit. It checks the wheel, runs tests against the installed distribution, publishes to PyPI, then publishes the same version to the MCP Registry. Both use GitHub OIDC; no local publishing token is required. If only the registry job fails, rerun that failed job rather than republishing the immutable PyPI version. When a workflow correction requires a new run, select registry_only to retry registry publication without uploading the existing PyPI version again.

Create the matching GitHub release from that same commit and attach the workflow's python-distributions artifact.

Available Tools

3 tools
compress_pdfCompress PdfC

Read a local PDF, run Mistral OCR, recompose it into dense packed PNG images, and create a retrievable compression job.

ParametersJSON Schema
NameRequiredDescriptionDefault
pdf_pathYes
inline_imagesNo
chars_per_imageNo

TDQS

C2.9/5.0
Behavior3/5

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

With no annotations, the description carries the full burden. It usefully discloses the OCR dependency (an external Mistral service, implying network/cost implications) and that the output is a retrievable job, but it omits permissions, cost/rate limits, whether existing output is overwritten, and synchronous vs. asynchronous behavior.

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?

A single front-loaded sentence listing the pipeline in execution order, with no filler. It is slightly dense but every clause conveys a distinct step.

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

Completeness2/5

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

For a tool with three undocumented parameters, no annotations, and no output schema, the description should explain the tuning knobs and how the returned job is consumed. It leaves the caller unable to set inline_images/chars_per_image meaningfully or know what the job handle is.

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

Parameters2/5

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

Schema description coverage is 0%, so the description must compensate for all three parameters, and it does not. pdf_path, inline_images, and chars_per_image are never named or explained; 'dense packed PNG images' only loosely gestures at the packing knobs without mapping to any parameter.

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

Purpose4/5

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

The description names a specific multi-step operation (read PDF, OCR, pack PNGs, create job) rather than restating the name 'compress_pdf'. It implicitly distinguishes itself from the retrieval siblings get_job_manifest and get_packed_images by being the creator of the job, though it never explicitly names them.

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

Usage Guidelines2/5

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

There is no explicit when-to-use guidance, no prerequisites (e.g., a running OCR service, local file access assumptions), and no mention of when to prefer the sibling retrieval tools. Usage is only inferable from the pipeline description.

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

get_job_manifestGet Job ManifestC

Load the saved manifest for a previously compressed PDF job.

ParametersJSON Schema
NameRequiredDescriptionDefault
job_idYes

TDQS

C2.9/5.0
Behavior2/5

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

No annotations are provided, so the description carries the full burden. 'Load' implies a read, but it says nothing about behavior if job_id is unknown or expired, whether the manifest persists across sessions, or what permissions are needed.

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?

A single front-loaded sentence with no filler. It is efficient, though its brevity borders on under-specification rather than tight editing.

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

Completeness2/5

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

With no annotations, no output schema, and an undocumented parameter, the description should explain what a 'manifest' contains and what failure looks like. It leaves the return shape and error semantics entirely unspecified.

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

Parameters3/5

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

Schema coverage is 0% and the single job_id property has no description, so the description must compensate. It does add one real semantic constraint — the job must be a previously compressed PDF job — but says nothing about the identifier's format or where the agent obtains it.

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

Purpose4/5

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

States a specific verb (Load) and resource (saved manifest) scoped to a previously compressed PDF job. An agent can distinguish this read of a prior job's artifact from the sibling compress_pdf write path, though it never names the siblings explicitly.

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

Usage Guidelines2/5

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

Only implied usage: the phrase 'previously compressed PDF job' hints that compress_pdf must have run first, but there is no explicit when-to-use, no when-not-to-use, and no mention of the alternative get_packed_images.

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

get_packed_imagesGet Packed ImagesC

Return one or more previously generated packed PNG images for a stored job.

ParametersJSON Schema
NameRequiredDescriptionDefault
limitNo
job_idYes
start_indexNo

TDQS

C2.7/5.0
Behavior2/5

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

No annotations are provided, so the description carries the full behavioral burden. It implies a read of existing output, but says nothing about what happens if the job_id is unknown, whether the images are returned as URLs, paths, or binary data, or how failure is surfaced.

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?

A single efficient sentence with the resource and target scope front-loaded. It is not padded, though it is arguably too terse to earn top marks.

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

Completeness2/5

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

With no annotations, no output schema, and three parameters at 0% description coverage, the definition is under-specified: it does not explain the return shape (one image vs many, and in what form) or how limit/start_index paginate results.

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

Parameters2/5

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

Schema description coverage is 0% and the description never mentions any of the three parameters. 'One or more' loosely gestures at limit, but limit, start_index, and their defaults/pagination semantics are entirely undocumented in both schema and description.

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

Purpose4/5

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

The description states a concrete verb ('Return') and resource ('packed PNG images'), plus the scope constraint that they are previously generated for a stored job. It is clearly distinguishable from compress_pdf and get_job_manifest, though it never names those siblings.

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

Usage Guidelines2/5

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

There is no explicit when-to-use guidance, no prerequisites (e.g. a job must have completed), and no mention of alternatives or when not to call it. The phrase 'for a stored job' only weakly implies the usage context.

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

Tool Schema Changelog

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

  1. 3 tool updatesv0.1.4
    • First observedcompress_pdf
    • First observedget_job_manifest
    • First observedget_packed_images
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

B3.2/5.0

Scored across 3 tools

Disambiguation5/5

Each tool has a distinct purpose: compress_pdf initiates processing, get_job_manifest retrieves metadata, and get_packed_images fetches image outputs. The verbs and objects clearly differentiate them.

Naming Consistency5/5

All tools follow a consistent verb_noun pattern (compress_pdf, get_job_manifest, get_packed_images) with snake_case, using standard verbs like 'compress' and 'get'.

Tool Count3/5

With only 3 tools, the set is minimal but covers the core lifecycle. However, it feels thin for a server that performs OCR, compression, and job management; additional tools like listing jobs or deleting outputs might be expected.

Completeness3/5

The tools cover creating a compression job, retrieving its manifest, and fetching packed images. Notable gaps include no tool to list all jobs, check job status, or delete a job, which limits lifecycle management.

Maintenance

ActivityMaintained
ResponsivenessNo issues

Resources

Related MCP Connectors

Related MCP Servers