Browse categories
Explore
Fiverr Pro
English
$
USD
I will evaluate ai text and audio in french and spanish
AI Language Evaluation, French and Spanish, Excel, Digital Content
Native French and Spanish speaker with C1 English. I provide human AI text and audio evaluation: TTS quality, linguistic QA, transcription, instruction-following and response comparison, drawing on pr...
About this Gig
Need a human quality check for AI-generated content in French or Spanish?
I am a native speaker of both languages with C1 English and project-based AI evaluation experience through Outlier. I review text or audio against up to 4 agreed criteria, including language quality, clarity, instruction-following, pronunciation, pacing and naturalness.
You receive a scored spreadsheet with specific issues and actionable findings. Standard adds prioritized findings; Premium adds a concise bilingual QA report when applicable.
Basic: up to 500 words OR 3 audio minutes, max 10 outputs.
Standard: up to 1500 words OR 10 audio minutes, max 30 outputs.
Premium: up to 3000 words OR 20 audio minutes, max 60 outputs across French and/or Spanish.
Limits apply to the full batch. Prompts and all A/B versions count toward the limit; one A/B pair counts as 2 outputs.
Packages cover evaluation onlynot rewriting, full transcription, translation, model training or voice generation. External factual research is excluded unless agreed through a custom offer.
Send an authorized sample, target locale and evaluation criteria before ordering. Extra volume, mixed requirements or additional criteria require a custom o
FAQ
Which languages and variants do you cover?
French and Spanish as a native speaker. Specify the target country and audience; I confirm regional requirements from your sample. Reports can be in English, French or Spanish. Basic and Standard cover one language; Premium can cover both within its total limit.
Can you compare two AI responses or TTS versions?
Yes. Both versions count toward the word/minute and output limits. One A/B pair is two outputs. Provide the original prompt or source text and the criteria for choosing a winner. All package limits apply together.
Does this include transcription, translation or rewriting?
No. Packages cover evaluation with scores and findings. Extra transcription, corrections, mixed text/audio batches and larger volumes require a custom offer agreed before ordering.
Do you verify factual accuracy?
I check consistency with references you provide and flag unsupported statements. External research and specialist validation are not included. This is a linguistic review, not scientific certification of a model.
Can you share work from private AI projects?
No. Platform and client work remains confidential. Examples must use independently created or properly licensed material. Send only authorized files, preferably anonymized, without private platform tasks or unnecessary personal data.
What counts as a revision?
A revision corrects or clarifies the evaluation of the original material using the agreed rubric. New outputs, another language, extra volume or a changed rubric are new work and need a custom offer.
