DeepMark

Create Exam Paper from Text

paper_create

Create a complete, markable exam paper from pasted text. Provide the question paper text (required) and, if the teacher has one, the mark scheme text — DeepMark extracts it. Omit the mark scheme to have DeepMark generate schemes (lower quality, but the paper is still markable). Send RAW text: do not pre-structure it into questions or author a mark scheme yourself. Extraction runs asynchronously — this returns a job_id. Poll ingestion_get until it reaches awaiting_confirmation, then call paper_confirm to mint the paper. RATE LIMITED: up to 3 papers per 24 hours per account (marking existing papers is unaffected).

Input Schema

TableJSON Schema

Name	Required	Description
`hints`	No	Optional cover-identity hints. SOFT grounding for extraction — they steer the model, they do not override the paper text.
`stimulus_text`	No	Optional source/stimulus material the questions refer to (a passage, dataset, case study). Supplied as extraction context; it is not itself turned into questions.
`mark_scheme_text`	No	RAW text of the mark scheme, if the teacher has one. Read jointly with the question paper so marks and numbering align. Send raw text, NOT a mark scheme you authored. Omit entirely when there is no mark scheme — DeepMark generates one per question (lower quality than a real scheme, but the paper is still markable).
`question_paper_text`	Yes	RAW text of the question paper — questions, bracketed marks, section headers, as printed or pasted. Send the source text and let DeepMark extract the questions. Do NOT pre-structure it into questions yourself.

Output Schema

TableJSON Schema

Name	Required	Description
`job_id`	Yes
`status`	Yes	The job's initial status (pending).
`poll_with`	Yes	The tool to poll this job with (ingestion_get).

TDQS

A4.7/5.0

Behavior5/5

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

Beyond the annotations (which only state this is not read-only), the description discloses that extraction runs asynchronously, returns a job_id, and is rate-limited to 3 papers per 24 hours. It also candidly notes the quality tradeoff when the mark scheme is omitted, adding meaningful behavioral context.

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

Conciseness5/5

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

The description is a single dense paragraph that front-loads the core action, then efficiently covers the workflow, raw-text caveats, and rate limits. Every sentence contributes necessary information with no redundancy, making it concise yet complete.

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?

For a complex tool with async behavior, a multi-step workflow, and rate limits, the description covers the entire lifecycle: input requirements, asynchronous job creation, polling, confirmation, and quality considerations. The existence of an output schema means return values need not be spelled out, and this description fills all contextual gaps.

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?

All 4 parameters are already covered by schema descriptions (100% coverage), so the description adds little parameter-level detail. It reinforces that the question paper text is required and explains the consequence of omitting the mark scheme, but those points are also present in the schema, keeping this at the baseline of 3.

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?

The description opens with 'Create a complete, markable exam paper from pasted text,' clearly stating the verb and resource. It distinguishes itself from sibling tools like paper_confirm by explaining that this tool returns a job_id and that paper_confirm mints the paper, so the purpose is unambiguous.

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

Usage Guidelines5/5

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

The description gives explicit when-to-use guidance: send RAW text, do not pre-structure it, and omit the mark scheme only if you want DeepMark to generate one. It also provides the full async workflow—poll ingestion_get until awaiting_confirmation, then call paper_confirm—and notes the rate limit, which differentiates it from marking tools.

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

All DeepMark tools Inspect

Try in Browser

RayrunBuild, deploy, and operate MCP servers without becoming an infrastructure team.ray.runAd

Glama MCP Gateway

Add one secure layer between your agents and this server.

TDQS

A4/5.0

Disambiguation4/5

The tools map well to distinct resources and workflow stages, and the async flows are clearly explained. The main overlap risk is paper_create/paper_confirm and markResult_get/submission_get, where names point at similar actions but the descriptions make the boundaries usable.

Naming Consistency4/5

Most names follow the same resource_action snake_case pattern, e.g. paper_get, submission_mark, markScheme_update. markResult_get breaks the pattern slightly by mixing camelCase into an otherwise snake_case set, but the naming stays readable and predictable overall.

Tool Count5/5

Ten tools is a well-scoped size for this domain: paper creation, confirmation, retrieval, question/mark-scheme correction, and submission grading are covered without excess granularity or obvious filler.

Completeness4/5

The main end-to-end workflow—create extraction job, confirm paper, mark submission, poll results—is complete and coherent. Minor gaps exist, such as no paper deletion, no submission listing, and no direct way to fetch a mark scheme separately, but they are not blocking for the core grading loop.

Resources

Need Help?