What are the opening hours?

Monday 9am - 5pm Tuesday 9am - 5pm Wednesday 9am - 5pm Thursday 9am - 5pm Friday 9am - 5pm

SadiGroup, Stella Court, Sunnyvale, CA (2026)

08/28/2026

VENDOR CALL | MULTI-COUNTRY EGOCENTRIC VIDEO COLLECTION – POC EVALUATION

SadiGroup invites experienced vendors/teams to submit capacity and pricing for a multi-country first-person (Ego/POV) video collection.

SCOPE
• Target: 5,000 accepted 10-minute clips, distributed evenly across the UK, Canada, Singapore, South Africa, New Zealand, Australia and India (~715/country; final allocation TBD).
• 3,000 clips (~429/country) is a comparison benchmark only, not guaranteed volume.
• Each asset is one continuous ~10-minute clip; 6 accepted clips = 1 valid footage hour.
• Vendors may support one or more countries but must state exact capacity per country. Delivery date: TBD; allocation is subject to quality/value review.

CAPTURE
• Insta360 X5 only; vendor must independently own, buy or rent devices.
• Head-mounted by Easy Clip/hat clip; horizontal only.
• FreeFrame, single lens, 4K 3840×2160, 30fps, MaxView 170°, 16:9; Horizon Lock/Jitter Blur Off; Standard filter; Matrix metering; EV 0; WB Auto.
• Authentic, unscripted day-in-life behavior. Each activity must begin and finish within the clip, normally with 2–4 natural sub-activities.
• 54 routines covering residential, commercial, urban/suburban/natural outdoor, arts, commuting and errands: cooking, cleaning, home/office work, café/restaurant visits, shopping, gardening, hiking, driving, etc.
• No fixed indoor/outdoor split; residential/outdoor prioritized, commercial where feasible.
• Target: 80% daytime/nominal (>100 lux), 20% sunset/night (

08/28/2026

🌎 Vendor Hiring | Latin American Spanish Voice Recording Project | 2,736.5 Valid Hours

SadiGroup is currently onboarding experienced vendors and organized teams for a large-scale Latin American Spanish ASR / Voice Recording Project.

The active scope covers Tasks 1, 3, and 4. Task 2 has been postponed for now.

Project Details:

Rate: USD 25 per accepted valid hour

Total Volume: 2,736.5 valid hours

Deadline: November 20, 2026

Current Country Allocation:

Colombia: 281.5 valid hours
Peru: 900 valid hours
Chile: 675 valid hours
Costa Rica: 880 valid hours

Task 1:
Sentence recording. Each participant may record a maximum of 0.5 valid hour.

Task 3:
Each participant records 525 sentences.

Task 4:
Sentence recording. Each participant may record a maximum of 0.5 valid hour.

Participant requirements include native speakers from the required countries, compliance with the required age distribution, and a 1:1 male-to-female ratio.

Selected vendors must have strong recruitment capacity, reliable project management, internal quality control, and the ability to meet country-specific and demographic quotas.

Delivery Commitment:

The complete 2,736.5 qualified valid hours must be delivered by November 20, 2026.

If the final qualified delivery is below 2,300 valid hours, a 30% deduction will apply.

If delivery is between 2,300 and 2,500 valid hours, a 20% deduction will apply.

If delivery is between 2,500 and 2,736.5 valid hours, a 10% deduction will apply.

Additional workload may be allocated based on quality, production capacity, and overall performance.

Interested vendors should contact us with:

Countries they can support
Estimated participant capacity
Expected valid hours per country
Weekly production capacity
Relevant speech data collection experience
Confirmation that USD 25 per accepted valid hour is workable

Email: [email protected]

SadiGroup
Building the Human Intelligence Behind AI.

08/27/2026

📢 VENDOR CALL | HEBREW & FRENCH AUDIO QA PROJECT

SadiGroup is seeking experienced vendors, agencies, and managed teams with native Hebrew and French (France & Canada) resources for an online speech-data quality assurance and annotation project.

Languages Required

• Hebrew
• French (France & Canada)

Project Overview

• Work mode: Fully remote
• Estimated volume: Approximately 400 hours
• Project duration: Expected to continue through the end of October 2026
• Current workload: Limited while data collection is underway
• Expected submission schedule: Two batches per week once annotation begins
• Estimated batch size: 100–300 sentences
• Monthly workload will be calculated according to the actual number of sentences checked
• Feedback on submitted results must be provided by the following day

Rates

• Sentence-level QA: USD 0.03 per checked sentence of 2–7 seconds
• Accent verification: USD 0.05 per recording of 15–30 minutes. Full listening is not required; reviewers only need to listen sufficiently to confirm the speaker’s accent
• Data annotation: USD 0.02 per sentence of approximately 4–8 seconds

Main QA Responsibilities

• Confirm that recordings contain natural conversational speech
• Reject formal reading, language-teaching content, strong or unsuitable accents, unsupported dialects, excessive background music, poor audio, synthetic voices, telephone or replayed recordings, and content mainly in another language
• Verify that source audio-visual segments are at least five minutes long and that no account/person contributes more than three hours
• Check source information, including the original URL, duration, program title, account name, six-digit file naming, and matching 16 kHz, 16-bit, mono WAV files
• Identify invalid speech involving major speaker overlap, unintelligible content, disruptive noise, frame loss, clipping, abnormal energy, non-human voices, television/radio audio, non-target language, or sensitive content
• Verify sentence-level segmentation, with a maximum duration of eight seconds and an ideal average of five to six seconds
• Ensure that each sentence contains only one speaker, boundaries are correctly positioned, 0.2–0.3 seconds of silence is preserved where possible, and pauses longer than two seconds are split correctly
• Verify speaker IDs and gender labels
• Ensure verbatim transcription with no missing, additional, or misspelled words
• Check capitalization, written-out numbers, spoken abbreviations, permitted punctuation, repetitions, interjections, unfinished words, and colloquial expressions
• Confirm correct use of project tags, including [N], [HM], [OVERLAP/]…[/OVERLAP], [IVS], [OIVS], and [PIL]

Vendor Requirements

• Native Hebrew or French (France & Canada) QA resources
• Proven experience in audio transcription, segmentation, annotation, or linguistic QA
• Ability to handle regular twice-weekly batches
• Capacity to provide corrections and feedback by the following day
• Dedicated project manager or team leader
• Strong internal quality-control process
• Ability to follow the complete project guidelines and acceptance requirements

This opportunity is intended for professional vendors, agencies, and managed teams.

To apply, please provide:

1. Company or team profile
2. Supported language(s)
3. Number of available native QA specialists
4. Relevant project and platform experience
5. Daily and weekly production capacity
6. Earliest available start date
7. Confirmation that all listed rates are acceptable

Contact Roben, SadiGroup Project Manager:
https://wa.me/message/ZEHDI6TQWLTVO1

SadiGroup
Building the Human Intelligence Behind AI.

08/25/2026

VENDOR & AGENCY CALL | Multilingual Speech Data Collection & Annotation

SadiGroup invites qualified vendors and agencies—no individuals—to support a multilingual natural-speech dataset for AI training.

LANGUAGES

Finnish, Greek, Hebrew, Ukrainian, Bulgarian, Swahili, Azerbaijani, Kazakh, Slovenian, Croatian, Serbian, Lithuanian, Latvian, Estonian, Persian, Afrikaans, Icelandic, Albanian, Catalan, Cebuano, Irish, Javanese and Uzbek.

SCOPE

• At least 200 accepted hours per language.
• Source public online videos; subtitled content is preferred.
• Standard national language variety; speakers aged 18–65.
• Natural/free-talk or professional-domain speech; single- and multi-speaker content accepted. Reading recordings and film/TV content are prohibited.
• Each source segment must exceed 5 minutes; maximum 2 hours from the same account/speaker.
• Add a domain label to every sample, e.g. finance, technology or gaming. Vertical-domain material must contain at least 60% professional terminology.
• Exclude strong accents, mostly non-target speech, synthetic/replayed/phone audio, persistent music, heavy noise and poor-quality audio.
• Deduplicate by MD5, URL and text similarity; edit-distance similarity must not exceed 60%.

DELIVERABLES

• Six-digit filenames plus ID, URL, language, accent, country, duration, category, program/show, account and title.
• Matching WAV audio: 16 kHz, 16-bit, mono, using the same filename.
• URL spreadsheet with at least 98% accuracy. Every audio must match a video and URL; audio under 15 KB is invalid.
• Staged delivery: 5% → 15% → 40% → 100%, with a self-inspection report at each stage. Proceed only after acceptance.

ANNOTATION

• Sentence-level, semantically coherent segments: maximum 8 seconds; recommended average 5–6 seconds.
• Retain 0.2–0.3 seconds of silence where available; no clipping or mixed speakers in one segment.
• Assign speaker ID and gender; transcribe verbatim using target-language conventions. Write numbers in words, avoid abbreviations, and apply correct punctuation, spacing and required noise/overlap/invalid/PII tags.
• Reject unresolved overlap, unintelligible/noisy or abnormal audio, non-target/synthetic speech, PII, and political, religious, pornographic or violent content.

QUALITY

• Word accuracy: at least 98%, including punctuation.
• Speaker-role accuracy: at least 90%.
• Special-label sentence accuracy: at least 90%.
• Native speakers preferred; QA inspectors must have long-term residence in the target-language country.

COMMERCIAL TERMS

• Collection: USD 2 per accepted valid hour. This covers public-web video sourcing/collection only.
• Annotation: Quote your rate per accepted valid hour.

TO APPLY

Send your supported languages, available volume per language, timeline, annotation rate, team profile, QA workflow, relevant experience and samples.

Contact Project Manager Roben:
https://wa.me/message/ZEHDI6TQWLTVO1

08/24/2026

📢 VENDOR CALL: American English Wake-Word Data Collection

SadiGroup is seeking qualified data collection vendors, agencies, and recruitment partners in the United States for an American English wake-word recording project.

Project scope:

• Total requirement: 550 unique native American English speakers
• Gender distribution: 275 male and 275 female speakers
• Age distribution:

* 18–45 years: 60%
* 46–60 years: 40%
* Balanced representation from Northern and Southern regions of the United States
* Each participant records 30 repetitions:
* 15 at normal speaking speed
* 15 at fast speaking speed
* Recording distance: approximately 20 cm from the mobile device
* Recording environment: quiet indoor space with no background noise or reverberation
* All participants must sign the required authorization form

Technical requirements:

• WAV format
• 16 kHz sample rate
• 16-bit, mono audio
• Clear, natural, standard American English pronunciation
• Signal-to-Noise Ratio above 30 dB
• Complete and accurately matched participant metadata

Important conditions:

• Participants must be real, unique individuals.
• Anyone who has already participated in this project through another vendor is not eligible.
• Vendors must conduct recruitment, participant verification, metadata management, recording supervision, and internal quality control.
• Rejected or incomplete recordings must be corrected or re-recorded.

💰 Rate: USD 20 per fully completed and accepted participant.

This opportunity is open to established vendors, agencies, and organized recruitment teams only. Individual applications will not be considered.

Interested vendors should send:

1. Company or team profile
2. Number of eligible participants available
3. Gender, age, and regional coverage
4. Expected recruitment and delivery timeline
5. Relevant speech or audio data collection experience
6. Quality-control methodology

Please contact SadiGroup or connect directly with our Project Manager:

https://wa.me/message/ZEHDI6TQWLTVO1

SadiGroup
Building the Human Intelligence Behind AI.

08/21/2026

📢 Vendor Call | Hebrew Audio Transcription & Tagging Project

SadiGroup is inviting qualified vendors, annotation agencies, and companies with managed teams of native Hebrew speakers to support a large-scale colloquial audio annotation project.

Project Details:

• Language: Hebrew
• Task: Audio transcription and tagging, including speaker identification and noise/non-speech tags
• Data type: Colloquial and conversational Hebrew video content
• Total volume: 500 valid audio hours
• Quality requirement: Minimum 98% word accuracy, subject to QA alignment
• Rate: USD 54 per accepted valid audio hour
• Completion deadline: September 30, 2026
• Platform: DataPlus

To apply, please provide:

1. The number of native Hebrew annotators currently available.
2. Confirmation of whether your team has previous experience working on the DataPlus platform.
3. Your expected weekly capacity in valid audio hours.
4. Details of your transcription, tagging, and internal QA experience.
5. Your earliest possible starting date.

This opportunity is intended for established vendors, agencies, and companies capable of recruiting, managing, and quality-controlling their own teams. Individual freelancer applications will not be considered.

Interested vendors may contact SadiGroup and submit their company profile, relevant experience, available capacity, and responses to the questions above.

08/19/2026

📢 GLOBAL VENDOR CALL / RFQ | MULTILINGUAL AUDIO DATA COLLECTION

SadiGroup invites qualified vendors, agencies, recording studios and companies with recruitment and QA teams to submit capacity and pricing for an upcoming multilingual audio program.

🚫 Vendors and companies only. Individuals and freelancers will not be considered.

PROJECT SCOPE

• Wake-word, command-word and speech-recognition text recordings
• Estimated scale: approximately 5,300 participants and 5,000 recording hours
• Wake/command tasks: children, adults and seniors, depending on language
• Speech-recognition tasks: adults
• Gender target: 1:1 male-to-female
• Slow, normal and fast speech; some tasks may require extremely fast speech
• Native speakers only, with mainstream regional pronunciation unless a designated accent is requested
• Final volume and allocation are subject to confirmed requirements

LANGUAGES & ACCENTS

Europe/Eurasia: Russian, French, German, Italian, Norwegian, Dutch, Polish, Swedish, Finnish, Danish, Greek, Czech, Swiss German, Estonian, Slovak, Slovenian, Bulgarian, Croatian, Lithuanian, Hungarian, Ukrainian and Turkish.

Asia: Japanese, Thai, Malay, Indonesian, Hindi, Vietnamese, Korean, Uzbek, Filipino and Kazakh.

MENA: Arabic, Persian, Hebrew and regional Arabic from Saudi Arabia, UAE, Qatar, Kuwait, Bahrain and Oman.

Spanish/Portuguese: Spanish, European Spanish (Spain), Mexican Spanish, Chilean Spanish, Peruvian Spanish, Portuguese, European Portuguese and Brazilian Portuguese.

English varieties: General, Indian, Pakistani-accented, Australian-accented and Singaporean-accented English; GCC regional English; and Scottish, Welsh, New York, Washington and other regional accents.

TECHNICAL & QUALITY REQUIREMENTS

• High-fidelity equipment or approved Apple-device recording applications
• Professional studio or quiet-bedroom environment; noise ≤200 smpl, without echo or interference
• 16 kHz, 16-bit, mono WAV; no noise reduction or amplitude normalization
• Voice amplitude: 5,000–25,000 smpl
• Clear, fluent, exact-script delivery with no omissions, additions or mispronunciations
• Accent screening before production and a minimum 98% acceptance rate
• Rejected recordings must be re-recorded within the required turnaround time

QUOTATION & APPLICATION

Submit USD-per-person rates using SadiGroup’s quotation template, separated by language, age group, task type and sentence-count band.

Include:

• Company profile, registration country and website
• Language and regional coverage
• Monthly participant and recording-hour capacity
• Recruitment, consent, accent-screening and QA processes
• Recording setup and relevant experience
• Pilot readiness and delivery schedule
• Completed quotation template

Send your proposal by DM to SadiGroup, addressed to Roben, Project Manager.

Subject: Multilingual Audio Collection – Vendor Proposal – [Company Name]

Please do not post quotations in the comments.

08/18/2026

📢 VENDOR & AGENCY CALL | TWO UPCOMING SOUTH KOREA DATA PROJECTS

SadiGroup is inviting qualified vendors, agencies, and service providers to express interest in two upcoming projects in South Korea.

These opportunities are exclusively for established vendors and agencies capable of providing and managing reliable teams. Individual applications will not be considered.

🔹 PROJECT 1: SOUTH KOREA VLA DATA ANNOTATION

We are building a reserve pool of annotators to support upcoming Vision-Language-Action (VLA) autonomous-driving data annotation projects.

Annotator requirements:

• Hold a valid South Korean driver’s license
• Have practical driving experience in South Korea
• Korean and Chinese candidates are both acceptable
• Be able to work through our internationally deployed annotation platform
• Be available for online work and, if required, onsite work in South Korea

Vendors should confirm:

1. Previous experience with autonomous-driving or VLA data annotation
2. Available team size and daily capacity
3. Ability to provide a stable and reliable team
4. Ability to support onsite annotation in South Korea
5. Supported onsite locations
6. Separate per-person, per-day rates for:
* Online annotation
* Onsite annotation
7. Expected team ramp-up time

🔹 PROJECT 2: OCR PERSONAL-DOCUMENT DATA TRANSCRIPTION

This project involves sensitive personal information. Korea-based service providers are strongly preferred.

Overseas service providers will be considered only if they hold valid ISMS-P certification.

Vendors should confirm:

1. Whether the company is a Korea-based/local service provider
2. If based outside Korea, whether the company holds valid ISMS-P certification
3. Previous experience with OCR data transcription, particularly personal or sensitive documents
4. Security, confidentiality, and data-protection procedures
5. Available team size and production capacity
6. Expected rate, including the pricing unit, currency, taxes, and services covered

📩 HOW TO APPLY

Please send the following by message to SadiGroup, addressed to Roben, Project Manager:

• Project 1, Project 2, or both
• Company profile and operating location
• Relevant project experience
• Available capacity
• Proposed rates
• Primary contact details

08/15/2026

📢 VENDOR & AGENCY CALL | Greek, Czech and Hebrew Video Data Project

SadiGroup is seeking qualified vendors and agencies for an upcoming colloquial video collection and audio annotation project covering:

🇬🇷 Greek
🇨🇿 Czech
Hebrew

This opportunity is strictly for established vendors or agencies with native-language teams, project-management capacity and a structured QA process. Individual applications will not be considered.

Project scope:

• Source and collect publicly available online videos containing natural, colloquial or spontaneous speech
• Extract and annotate the corresponding audio
• Each video must feature at least two speakers
• Each video segment must be at least five minutes long
• Data from the same person or account must not exceed two hours
• Domains are currently flexible
• Videos must have clear, intelligible speech and manageable background noise
• Formal reading, synthesized voices, prolonged music, poor-quality audio and sensitive, political, religious, violent or adult content must be excluded
• Vendors must ensure lawful sourcing and retain the original URL and source metadata

Annotation includes sentence-level segmentation, timestamps, verbatim transcription, speaker identification, gender information, required audio-quality tags and metadata preparation. Extracted audio should be delivered as 16 kHz, 16-bit, mono WAV files with the corresponding TXT and metadata files.

Interested vendors should provide the following information separately for Greek, Czech and Hebrew:

1. Estimated number of valid video hours your team can collect
2. Number of qualified native annotators available
3. Rate per accepted valid hour for collection and annotation, indicating whether QA is included
4. Weekly capacity and earliest possible start date
5. Relevant experience in colloquial or spontaneous-speech annotation
6. Confirmation that your team can comply with the sourcing, privacy and content requirements

Vendors may apply for one or multiple languages.

Please contact Roben, SadiGroup Project Manager, with your company profile and detailed proposal. Include “Greek/Czech/Hebrew Video Project – Vendor Application” in the subject line.

SadiGroup — Building the Human Intelligence Behind AI.

08/14/2026

SadiHub is now live — and we are inviting expert vendor companies to join our global delivery network.

SadiHub connects qualified vendors with projects aligned to their expertise, languages, geography, capacity and performance.

We are seeking experienced companies and specialist teams in:

• Data collection and annotation
• Transcription and quality assurance
• Localization and multilingual operations
• Human evaluation and RLHF
• Voice and speech-data production
• BPO and field operations

Approved vendors can build a verified profile, access suitable project opportunities, receive clear assignments, report progress, submit work for QA and develop a measurable performance history—all in one platform.

We are looking for reliable partners who can demonstrate quality, security, responsiveness and delivery capacity.

Apply to join SadiHub:

https://app.sadigroup.co

SadiHub — Building the Human Intelligence Behind AI.

SadiGroup

08/28/2026

08/28/2026

08/27/2026

08/25/2026

08/24/2026

08/21/2026

08/19/2026

08/18/2026

08/15/2026

08/14/2026

Address

Opening Hours

Telephone

Website

Alerts

Contact The Business

Shortcuts

Share

Category

Listing takedown request

Monday	9am - 5pm
Tuesday	9am - 5pm
Wednesday	9am - 5pm
Thursday	9am - 5pm
Friday	9am - 5pm