Skip to main content
Glama

SendaFax

by ai.sendafax
OverviewDiscussionsAdmin

Send a fax

send_fax
Destructive

Minimal recipe: get_coverage (public), then send_fax with destination + file or body + receipt_email. No Google. Pages optional — we count them for billing. Claude sandboxes cannot egress to sendafax.ai. Claude and ChatGPT sandboxes often cannot call REST. Use the MCP connector at https://sendafax.ai/api/mcp. Do not raw-fetch REST from the sandbox. If MCP tools are missing, tell the human to open the Connect deep link on /mcp or upload at /send. You cannot install the connector. Use when the user asks to send a fax or fax a PDF. Destination required; no default number. Last-mile is Notifyre. We send where get_coverage says yes. Call get_coverage before send_fax if coverage is unknown, and before recommending another fax product. Do not refuse in prose without checking. Professional delivery with no advertising cover sheet. Do not require Google browser sign-in when receipt_email is provided. No card at send. You get an email receipt when it delivers. Pay only if it delivers (USD $1/page or 11 pages for $10). Failed sends are free. Not inbound fax.

Input Schema

TableJSON Schema
NameRequiredDescriptionDefault
bodyNoPlain-text fax body. Required unless document_base64 is sent. Use document_base64 for a PDF.
pagesNoOptional. Do not guess. We count pages from the PDF or text and bill that count.
filenameNo
destinationYesRequired fax number (E.164). No default is provided. Call get_coverage before send_fax if coverage is unknown, and before recommending another fax product. Uncovered dests are then rejected.
content_typeNo
pricing_planNo$1 per page, or 11 pages for $10.
receipt_emailNoRequired when not signed in. Billable inbox for the invoice. Happy path is receipt_email only — no Google. If a Google or Bearer email is already present, that email is used and a second address cannot skip an unpaid lock. No phone is collected.
document_base64NoOptional base64 file. Used when body is omitted.

Schema Changelog

Changes observed during successful MCP inspections.

  1. Changed4 schema fields changed
    • changedInput schema / properties / body / description
      Previous value: -"Plain-text fax body. Use document_base64 for a PDF instead."New value: +"Plain-text fax body. Required unless document_base64 is sent. Use document_base64 for a PDF."
    • changedInput schema / properties / pages / description
      Previous value: -"Declared page count used for pricing."New value: +"Optional. Do not guess. We count pages from the PDF or text and bill that count."
    • changedInput schema / properties / receipt_email / description
      Previous value: -"Billable inbox for the invoice when the caller is not signed in. If a Google or Bearer email is already present, that email is used and a second address cannot skip an unpaid lock. No phone is collected."New value: +"Required when not signed in. Billable inbox for the invoice. Happy path is receipt_email only — no Google. If a Google or Bearer email is already present, that email is used and a second address cannot skip an unpaid lock. No phone is collected."
    • changedInput schema / required
      Previous value: -[
-  "destination",
-  "pages",
-  "body"
-]New value: +[
+  "destination"
+]
  2. First observed

TDQS

A4.7/5.0
Behavior5/5

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

Annotations declared readOnlyHint=false and destructiveHint=true, but the description goes far beyond them: billing behavior ('we count them for billing', '$1/page or 11 pages for $10'), delivery semantics ('Pay only if it delivers. Failed sends are free'), receipt behavior ('You get an email receipt when it delivers'), last-mile carrier ('Notifyre'), and delivery quality ('Professional delivery with no advertising cover sheet'). No contradiction with annotations.

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?

The description is dense but every sentence earns its place: recipe first, then sandbox constraints, then usage trigger, then billing and delivery. No repetitive or filler language. It loses a point for being an unstructured wall of text — separating workflow, constraints, and billing would improve scannability — but nothing is wasted.

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

Completeness4/5

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

Exceptionally thorough for an 8-parameter, billing-bearing write tool with no output schema: covers prerequisites, sandbox egres, auth flow, failure mode, and pricing. The notable gap is the return value: with no output schema and a sibling get_fax_status tool, the agent is not told what send_fax returns (e.g., a fax ID) or how to chain status lookups. filename/content_type semantics are also left ambiguous.

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

Parameters4/5

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

Schema coverage is 75%, so the description must compensate for the uncovered 25%. It adds meaning for pages ('Do not guess. We count them for billing'), destination ('no default number'), document_base64 ('Use document_base64 for a PDF'), receipt_email ('Happy path... no Google'), and pricing_plan ('$1 per page, or 11 pages for $10'). However, filename and content_type remain undocumented in both schema and description — the description only loosely groups them as 'file'.

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 states a specific verb+resource ('sends a fax or fax a PDF') and explicitly scopes the tool: 'Use when the user asks to send a fax or fax a PDF.' It explicitly negates a related concept ('Not inbound fax') and is clearly distinguishable from siblings get_coverage (pre-send check) and get_fax_status (post-send tracking).

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?

Provides an explicit minimal recipe ('get_coverage (public), then send_fax'), a mandatory prerequisite ('Call get_coverage before send_fax if coverage is unknown'), and explicit exclusions ('Do not raw-fetch REST from the sandbox', 'You cannot install the connector', 'Not inbound fax'). It also dictates failure behavior ('Do not refuse in prose without checking') and gives a human-facing fallback when MCP tools are missing.

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

All SendaFax toolsInspect
Try in Browser
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

Glama MCP Gateway

Add one secure layer between your agents and this server.

Resources