MPPtempoquality 0.52

Perplexity contextualized embeddings via pay-per-call MPP on Tempo L2

Price
per_call
Protocol
mpp
Verified
no
Endpoint
https://perplexity.mpp.paywithlocus.com/perplexity/context-embed

What it does

This endpoint provides access to Perplexity's contextualized embedding models through the Locus MPP (Micropayment Protocol), settling payments on Tempo L2 in pathUSD. Unlike standard embeddings, contextualized embeddings accept a full parent document alongside an array of text chunks, producing embeddings that are informed by the broader document context. This can improve retrieval quality for chunked documents in RAG pipelines.

Two model sizes are available: `pplx-embed-context-v1-0.6b` (0.6 billion parameters) and `pplx-embed-context-v1-4b` (4 billion parameters). The endpoint requires three fields: the model identifier, the full document text, and an array of chunks to embed. Pricing is approximately $0.001 per call according to the OpenAPI spec's payment info. Payment uses the MPP charge intent, meaning each request is a one-shot payment.

The endpoint is served at `https://perplexity.mpp.paywithlocus.com/perplexity/context-embed` as a POST request. Note that the probe did not receive a 402 challenge on HEAD/GET — this is expected since the endpoint only accepts POST. The root domain returns a helpful 404 directing callers to the skill file and llms.txt for documentation. Additional Perplexity endpoints (Sonar chat, web search, standard embeddings) are available on the same gateway.

Capabilities

contextualized-embeddingsdocument-aware-chunked-embeddingsperplexity-pplx-embed-contextmpp-chargetempo-l2-settlementpay-per-callrag-optimized-embeddings

Use cases

  • —Generating context-aware embeddings for document chunks in RAG pipelines
  • —Improving semantic search quality by embedding chunks with their parent document context
  • —Building vector databases where chunk embeddings reflect full-document meaning
  • —Comparing document sections with richer contextual representations

Fit

Best for

  • —RAG systems that chunk long documents and need context-preserving embeddings
  • —Agents that need pay-per-call embedding access without API key management
  • —Applications where chunk-level embeddings benefit from document-level context

Not for

  • —Embedding standalone short texts with no parent document (use the standard /embed endpoint instead)
  • —High-volume batch embedding jobs where per-call micropayments add overhead
  • —Use cases requiring non-Perplexity embedding models

Quick start

curl -X POST https://perplexity.mpp.paywithlocus.com/perplexity/context-embed \
  -H "Content-Type: application/json" \
  -d '{
    "model": "pplx-embed-context-v1-0.6b",
    "document": "The full text of your parent document goes here.",
    "chunks": ["First chunk of text.", "Second chunk of text."]
  }'

Example

Request

{
  "model": "pplx-embed-context-v1-0.6b",
  "chunks": [
    "Perplexity AI is a conversational search engine.",
    "It offers multiple model tiers including Sonar and Sonar Pro."
  ],
  "document": "Perplexity AI is a conversational search engine that uses large language models to provide answers grounded in real-time web results. It offers multiple model tiers including Sonar and Sonar Pro."
}

Endpoint

URLhttps://perplexity.mpp.paywithlocus.com/perplexity/context-embed
Transporthttp
Protocolmpp
CurrencypathUSD

Quality

0.52/ 1.00

The OpenAPI spec provides a clear schema with model names, required fields, and approximate pricing. However, the probe did not capture a live 402 challenge (only HEAD/GET were tried on a POST-only endpoint), no example response is available, and crawled pages all returned 404 with no additional documentation. The endpoint is likely live but unconfirmed via probe.

Warnings

  • —Probe did not receive a 402 challenge — HEAD and GET returned 404; the endpoint likely only responds to POST, so liveness is not confirmed by probe
  • —No example response schema or sample output is available in the provided material
  • —The 'chunks' field is typed as 'string' in the OpenAPI spec but described as 'Array of text chunks' — actual type is likely an array of strings
  • —Approximate pricing (~$0.001) is stated in the OpenAPI x-payment-info but exact amount is null

Citations

Provenance

Indexed frommpp_dev
Enriched2026-04-19 17:35:23Z · anthropic/claude-opus-4.6 · v2
First seen2026-04-18
Last seen2026-04-22

Agent access

JSONhttps://clawmart.sh/api/listings/pM5znY
Perplexity contextualized embeddings via pay-per-call MPP on Tempo L2 — Clawmart · Clawmart