1. Forested headwater streams are generally considered to be light‐limited ecosystems where primary production is reduced, and the main source of energy and nutrients is composed of allochthonous detritus. We hypothesised that in these ecosystems, the development of primary producers might also be limited by (1) competition for nutrients with leaf‐litter decomposers (e.g. bacteria and fungi), and (2) leaf‐litter leachates or allelopathic compounds produced by aquatic fungi. 2. To test these hypotheses, a 48‐day mesocosm experiment was performed in 12 artificial streams containing stream water inoculated with epilithic biofilm suspensions collected from a forested headwater stream. Three different treatments were applied: control without leaf litter (C), microbially conditioned leaf litter added at the beginning of the experiment and left to decompose throughout the experiment (L), or leaf litter renewed three times during the experiment (RL). 3. We predicted that (1) the presence of litter, through microbial nutrient immobilisation and allelopathy, would reduce primary production and that (2) this effect would be amplified by litter renewal. We also predicted that nutrient competition would mean that (3) leaf‐litter decomposers will alter primary producer community composition and physiology. These predictions were tested by analysing biofilm development, physiology, stoichiometry, and benthic algal community structure. To distinguish between the effects of nutrient immobilisation and allelopathy, the biofilm responses to leaf‐litter leachates collected after different microbial conditioning durations were also measured in a parallel laboratory experiment. 4. Contrary to our expectations, by day 28, primary producer growth was higher in the mesocosms containing leaf litter (L and RL) despite the rapid decrease in dissolved nutrients when leaf litter was present. After 48 days, the lowest phototrophic biofilm development was observed when leaf litter was renewed (RL), whereas phototrophic biofilm development was similar in the C and L treatments. Biofilm stoichiometry indicated that this effect was most probably related to greater nitrogen limitation in the RL treatment. The presence of leaf litter also affected primary producers' photophysiology, which could be attributed to changes in taxonomic composition and to physiological adjustments of primary producers. 5. Laboratory measurements showed that despite a strong inhibition of primary producer growth by unconditioned leaf‐litter leachates, microbially conditioned leaf litter had either low or no effects on the development of primary producers. 6. These results reveal that leaf‐litter decomposers can have both positive and negative effects on primary producers underlining the need to consider microbial interactions when investigating the functioning of forested headwater streams.