Paxa Labs vs BoostedTravel
A side-by-side look at the Paxa Labs and BoostedTravel MCP servers: transport, auth, tools, quality signals, and the config to install either one in Claude, Cursor, Windsurf or VS Code.
Verdict: choose Paxa Labs for text-to-speech work as a local process; choose BoostedTravel for travel & transportation as a local process. You can also run both side by side in one mcpServers block.
Which MCP server should you choose?
Choose Paxa Labs when
- You need dedicated capabilities in the Text-to-Speech domain.
- You prefer a local stdio subprocess transport.
- You have the keys it needs:
PAXA_API_KEY. - You want tools like
speak, queue_speech, control_playback.
Choose BoostedTravel when
- You need dedicated capabilities in the Travel & Transportation domain.
- You prefer a local stdio subprocess transport.
- Its auth model fits you: API key required.
- You have the keys it needs:
LETSFG_BEARER_TOKEN, LETSFG_API_KEY. - You want tools like
search_flights, resolve_location, unlock_flight_offer.
Feature & specification comparison
Rows where the two servers differ are marked.
| Spec | ||
|---|---|---|
| Category | Text-to-Speech | Travel & Transportation |
| Quality signal | 64/100 · Good | 72/100 · Great |
| Install command | Confirmed | Confirmed |
| Transport | Local subprocess (stdio) | Local subprocess (stdio) |
| Auth | Not declared | API key required |
| Pricing | Not declared | Free |
| Required env vars | PAXA_API_KEY | LETSFG_BEARER_TOKENLETSFG_API_KEY |
| Verified owner | Community listing | Community listing |
| GitHub stars | 10 | 2,188 |
| Directory activity | 1 views 0 installs 0 upvotes | 9 views 1 installs 0 upvotes |
| Tools | 14 | 14 |
| Full listing | View Paxa Labs | View BoostedTravel |
Tools & capabilities
Paxa Labs tools 14
speakSynthesize a short line with Paxa TTS and play it through this machine's speakers. Blocks until playback finishes. Waits for the currently playing audio but jumps ahead of queued long-form segments. For long content use queue_speech. Costs 10 credits per 1000 characters. Direction tags (paxa-tts-flash-v2, every voice): words in square brackets written into the text set the delivery from that point on, in English, Thai, or Chinese, e.g. [whispering], [speaking slowly, close to the microphone], [sad:0.5], or a blend [happy:0.7+sad:0.3]; [neutral] returns to the voice's own delivery. Sounds such as [laughs], [sighs], [gasps] play where written; [um] and [hmm] are spoken; <emph>...</emph> and <whisper>...</whisper> mark a span. Nothing in brackets is spoken, tag characters bill as text, and a tag the model cannot do is refused before any charge. On paxa-tts-flash-v1 only khanomchan and kaprao read tags, from a fixed emotion list.queue_speechRead long content aloud (stories, articles, books). Splits the text into segments, synthesizes ahead while playing, and returns immediately; segments play in order after anything already queued. Monitor and control with the control_playback tool. Costs 10 credits per 1000 characters. Direction tags (paxa-tts-flash-v2, every voice): words in square brackets written into the text set the delivery from that point on, in English, Thai, or Chinese, e.g. [whispering], [speaking slowly, close to the microphone], [sad:0.5], or a blend [happy:0.7+sad:0.3]; [neutral] returns to the voice's own delivery. Sounds such as [laughs], [sighs], [gasps] play where written; [um] and [hmm] are spoken; <emph>...</emph> and <whisper>...</whisper> mark a span. Nothing in brackets is spoken, tag characters bill as text, and a tag the model cannot do is refused before any charge. On paxa-tts-flash-v1 only khanomchan and kaprao read tags, from a fixed emotion list.control_playbackInspect and control the shared audio queue. status reports what is playing, what is queued, and any synthesis failures. pause and resume halt and continue playback (not supported on Windows). skip drops the current segment. clear stops playback and empties the queue.play_audioPlay a local audio file through this machine's speakers, for example a file produced earlier by text_to_speech. Blocks until playback finishes. Free (no API call).text_to_speechSynthesize speech with Paxa TTS and save it as an audio file without playing it: voiceovers, narration, audiobooks. Text beyond 5000 characters is split at paragraphs and sentences and the parts are joined into one file (needs ffmpeg). With subtitles set, an srt or vtt caption file that matches the voice word for word is written next to the audio. Costs 10 credits per 1000 characters, 12.5 with subtitles. Direction tags (paxa-tts-flash-v2, every voice): words in square brackets written into the text set the delivery from that point on, in English, Thai, or Chinese, e.g. [whispering], [speaking slowly, close to the microphone], [sad:0.5], or a blend [happy:0.7+sad:0.3]; [neutral] returns to the voice's own delivery. Sounds such as [laughs], [sighs], [gasps] play where written; [um] and [hmm] are spoken; <emph>...</emph> and <whisper>...</whisper> mark a span. Nothing in brackets is spoken, tag characters bill as text, and a tag the model cannot do is refused before any charge. On paxa-tts-flash-v1 only khanomchan and kaprao read tags, from a fixed emotion list.translate_to_thaiTranslate text from any language into Thai with Paxa Translate. Accepts a single string or up to 200 segments that share context. Costs 25 credits per 1000 text characters (minimum 2 credits); context and glossary characters bill at 8 per 1000.ocr_documentRun Paxa OCR on a local PDF, PNG, JPEG, or WebP file and return its content as GitHub-flavored Markdown (or structured blocks). Up to 50 pages and 10 MiB per file. Costs 6.5 credits per page.extract_fieldsExtract typed fields from a local PDF, PNG, JPEG, or WebP document (up to 20 pages and 10 MiB) with Paxa Document Extraction. You write a schema of fields; every value returned is text printed in the document, read as the field's type, or null with the reason. Nothing is inferred or looked up. Built for Thai documents (receipts, tax invoices, forms, IDs) and fine on English ones. Costs 13 credits per page for a schema of up to 50 leaf fields and 19.5 per page for 51 to 100; the page reading and the evidence spans cost nothing extra. A schema the API refuses costs nothing and names the field and reason. The schema parameter summarizes the dialect; https://paxalabs.com/docs/extraction has every type and key.transcribe_audioTranscribe a local recording or video of speech, Thai or any other language, code-switching included (mp3, wav, flac, ogg, m4a, aac, webm; or mp4, mov, mkv, avi and other video files, whose sound track is extracted locally with ffmpeg; up to 60 minutes and 25 MiB) with Paxa STT. The transcript comes back inline, so nothing needs to be read from disk afterwards; optionally it is also saved as txt, json (with word timings and segments), srt, or vtt, named after the recording. Existing files are never overwritten. Detects the language, handles code-switching, and can label speakers. Terms configured in PAXA_VOCABULARY or PAXA_VOCABULARY_FILE are pinned on every call. Costs 8.33 credits per minute of audio (0.1 minimum); subtitles and word timings cost nothing extra. Long recordings take minutes to process.listenRecord one spoken turn from this machine's microphone and return its transcript (Thai, English, or other languages, code-switching included). Paxa's realtime transcription detects when the user starts and stops talking, so the call returns as soon as the turn ends; the audio goes straight to the API and nothing is written to disk. Waits for any speech this server is playing to finish first. Use it after speak when you asked the user a question and want their spoken answer. Returns an empty transcript when nobody spoke within wait_seconds. Costs 12.5 credits per minute of audio sent, silence included (a few seconds of listening costs well under one credit). Needs a microphone recorder on this machine (ffmpeg or sox; parecord or arecord on Linux) and Node 22 or newer; the host app must have microphone permission.send_feedbackReport a wrong or poor output, an API problem, or a platform issue to the Paxa Labs team, tied to the request ids the other tools print ("Request id: ..."). A rating alone on at least one request id is a complete report. For a wrong output, add an example with the exact input, the output received, and the output expected. Free, no credits. Use it when the user says an output was wrong, or asks to report a problem; do not file reports the user did not ask for.list_voicesList the Paxa TTS voice roster: ids, names, gender, language, and character notes. Free.- +2 more on the listing page
BoostedTravel tools 14
search_flightsSearch hundreds of airlines for live flight prices — completely FREE, unlimited, read-only. Returns structured offers with prices, airlines, times, durations, and stopovers. Some offers carry `starlink` for in-flight Starlink Wi-Fi: "confirmed_all" / "confirmed_some" mean the carrier has FULLY fitted that aircraft type, "likely_all" / "likely_some" mean the rollout on that type is underway but incomplete. State only "confirmed_*" as fact; describe "likely_*" as "the airline is fitting this aircraft type, not guaranteed on your flight". Anything ending in "_some" has at least one leg WITHOUT it. An absent field means no information, NOT an absence of Wi-Fi. Covers airlines across all continents including low-cost carriers. Search is async: this tool polls for you, including waiting out the late split-ticket merge. Some offers are SPLIT TICKETS: two separately-issued tickets through a hub, each leg booked from whatever is cheapest for it (usually two different airlines), because no one seller offers the combination as a single ticket. They carry `split_ticket: "true"`, `combo_type: "virtual_interlining"` and `self_transfer: "unprotected"`. ALWAYS tell the user when an offer is a split ticket and what unprotected means: the tickets are not linked, so if the first flight is late and the connection is missed, the second airline owes nothing — no rebooking, no refund. Never present a split ticket as though it were one through-fare. Requires LETSFG_BEARER_TOKEN or LETSFG_API_KEY. See letsfg://guide resource for the full authenticate->search->book workflow.resolve_locationConvert a city/airport name to IATA codes. Always call before search_flights if you only have a city name. Read-only, safe to call multiple times.unlock_flight_offer[Developer API only] Confirm live price with the airline and reserve the offer for 30 minutes. NOT part of the agent flow and NOT needed before book_flight. It requires LETSFG_API_KEY (the paid, prepaid Developer API) and refuses to run on a Bearer token, because the PFS unlock endpoint does not exist — calling it that way used to 404. If you authenticated with `letsfg auth`, go straight from search_flights to book_flight. Requires a Developer API key. Legacy path — not idempotent.book_flightBook a flight from a search result. FLOW: authenticate (once) -> search_flights -> book_flight CHARGES: nothing from LetsFG. A completed booking pays the airline prices from airlines and the major booking sites. RESULT: either {"booked": true, "order_id": "..."} or {"booked": false, "booking_url": "..."} — the second means the booking genuinely did not complete and nothing was charged. That is a normal outcome, NOT a transient error: do not retry, give the user the booking_url. SAFETY: use REAL passenger details — names must match passport, email receives the e-ticket.get_flight_bookingPoll a flight booking started by book_flight. REQUIRED to learn the outcome: on a PFS Bearer token book_flight returns a booking_ref and state "booking_in_progress", not a PNR - the booking itself takes 4-11 minutes. Call it every 20-30 s with that booking_ref until state is terminal: completed -> PNR issued, card charged failed -> the hold was released, nothing was charged needs_attention -> a human is looking at it; do NOT rebook Do not rebook while the state is still moving, and do not treat a slow poll as a failure - the money is HELD, not taken, until the airline confirms. Refs last one hour past the booking start.resolve_hotel_cityConvert a place name to the supplier city id that search_hotels needs. Always call this first if you only have a city name. Read-only and safe to repeat. Use `Id` from the first result as city_id and `Name` as city_name.search_hotelsSearch real, bookable hotel inventory. Requires a payment method on file — the SAME card that authorises flight booking. That applies to search too, not just booking, because a search opens a real session at the supplier. Only free-cancellation, pay-later rates are returned, so everything you see can actually be booked on these terms. The result set is smaller than a metasearch and that is deliberate. `price` is what the guest pays. Keep `session_id` and the chosen offer's `combination_id_v2` — together they identify that exact rate, and book_hotel needs both. Takes up to a few minutes.book_hotelBook one hotel rate. Charges 5% of the price to the card on file immediately as a NON-REFUNDABLE reservation fee; the balance is paid directly to the supplier through the pay link we return, by balance_due_by (the supplier's own auto-cancellation date). Returns a booking_job_id, NOT the booking — a booking takes minutes. Poll get_hotel_booking until status is succeeded or failed. The fee is charged BEFORE the room is committed, so a declined card costs nothing: no reservation exists and nothing is charged. Send expected_price and expected_balance back exactly as search returned them. NOT idempotent — calling twice for the same rate books the room twice and charges two fees.get_hotel_bookingCollect the result of a booking started with book_hotel. Poll every ~20s. status is in_progress, succeeded or failed. On success you get confirmation, reservation_fee_charged, pay_link, balance_due, balance_due_by and the full cancellation ladder. Read-only and safe to repeat.cancel_hotel_bookingRelease a hotel reservation. Free until balance_due_by; after that the hotel's own ladder applies and can reach 100%. The ladder is in the booking terms, so check the cost first. The 5% reservation fee is NOT refunded. Takes over a minute; if it times out do NOT assume it failed — re-check before retrying.authenticateExplain how to connect a card so this server can search and book. Nothing is charged to connect — a 0.00 Revolut setup that saves the card so a booking can be charged later. Call with no arguments. It returns the current instructions and add_card_url (https://letsfg.co/connect). A PERSON must approve once in a browser — there is no endpoint that mints a token from card details, so do not ask the user for card numbers and do not try to automate this step. Two ways in: (a) add LetsFG as a connector in an assistant that supports remote MCP servers and approve it, or (b) any OAuth-capable client can register itself — see https://letsfg.co/for-agents, section "Option B". Both land on the same card screen. RETIRED 2026-09-02: the Stripe lanes (setup_url, setup_session_id, payment_method_id, card_token) and every token they issued. Passing them now fails. This does NOT create a Developer API billing account. Do not use setup_payment for this.setup_payment[Developer API only — you almost certainly want `authenticate` instead] Attaches a card to a PAID prepaid Developer API account. Refuses to run unless LETSFG_API_KEY is set, because agents kept calling this and creating billing accounts they did not need.- +2 more on the listing page
Ready-to-paste client configs
Paste either block (or both) into your client config, e.g. claude_desktop_config.json or ~/.cursor/mcp.json. We only show a config when the install command is confirmed.
Paxa Labs
{
"mcpServers": {
"paxa-labs": {
"command": "npx",
"args": [
"-y",
"@paxalabs/mcp"
],
"env": {
"PAXA_API_KEY": "YOUR_PAXA_API_KEY_HERE"
}
}
}
}
BoostedTravel
{
"mcpServers": {
"boostedtravel": {
"command": "npx",
"args": [
"-y",
"letsfg-mcp"
],
"env": {
"LETSFG_BEARER_TOKEN": "YOUR_LETSFG_BEARER_TOKEN_HERE",
"LETSFG_API_KEY": "YOUR_LETSFG_API_KEY_HERE"
}
}
}
}
Frequently asked questions
Paxa Labs is categorized under Text-to-Speech and uses a local stdio subprocess. BoostedTravel belongs to Travel & Transportation and uses a local stdio subprocess. Pick Paxa Labs for text-to-speech work and BoostedTravel for travel & transportation.
Copy the mcpServers block from the config section on this page into your client's config (for example claude_desktop_config.json or ~/.cursor/mcp.json), then fully restart the client.
Paxa Labs has no declared pricing — check the project's own site. BoostedTravel is listed as free. Auth: api key required. If a server needs environment variables such as API keys, define them under the env key in your client's MCP config.
Yes. MCP clients support multiple servers: add paxa-labs and boostedtravel as separate keys in the same mcpServers object, and your assistant will call tools from either one.
On AllMCPs, Paxa Labs has a quality signal of 64/100 (Good) with 1 views, 0 installs and 10 GitHub stars. BoostedTravel has 72/100 (Great) with 9 views, 1 installs and 2,188 GitHub stars.