Sponsored
TokenReply
TokenReply is an API gateway built for long AI roleplay chats. One key and one OpenAI-compatible endpoint (https://api.tokenreply.com/v1) reach DeepSeek, GLM, Claude, Gemini, Grok, Kimi and more, with setup guides for JanitorAI and SillyTavern. Prompt caching makes long chats cheaper, and automatic parameter clamping keeps roleplay frontends from hitting 400 errors. Free access: every account can call models marked Free on the models page (no time limit, rate limited), plus up to 40 calls per day across a rotating Models of the week selection; a daily reward adds 10 more calls that day. Plus is $5/month for higher limits. Submitted by the TokenReply team (vendor).
- Access: Permanent free access
- Models: Free models listed live at https://www.tokenreply.com/guides/free-ai-api (34 on 2026-10-23), e.g. big-pickle-free big-pickle-thinking-free gemma-4-31b-it google/gemma-4-31b-it dots-3-note-preview Models of the week (40 calls/day): rotating DeepSeek, GLM, Kimi, Gemini Flash, Grok, MiniMax models
- RPM: Unknown
- TPM: Unknown
- RPD: Unknown
- Resets: daily (Models of the week: 40 calls/day, +10 with the daily reward)
- Phone verification: No
- Card: No
- Deposit: No